Document MGZ5VMr9x2kQBDON78w75nbXL

Wri EmuMon PVC Trends 7 rends in suspension PVC manufacture m dynamic technology, largo reactors, zatlon, clean wall systems, and VCM Inmant, PVC should hold Its position a feeding thermoplastic Cameron, Conoco Chemicals Co., Houston; . Luncfean and J.H. McCuilty, Ur., (Conoco, hit., City, Okly. U-SflS of PVC ini' divided imo the broad ureas oi' -xiUe (or plasticized) and rigid (non-piasiicized) up- tions. Historically, flexible applications appeared >. Initially. plastici/ed PVC found use* ;n wire inxiUt- Slum ahei. calendered PVC. introduced as a "cr replacement. became widely known as "vinyl" (he public. Flexible PVC was initially processed on uipincul used in tlx- rithhcrimlnsiry. Now, however, xiali/ed equipment lias Ijceit developed lor flexible C. Rigid PVC. used in pipe. linings and building pim- luis grown rapidly in die past decade (Fig. I). proved resin qualiiy, stabilizer systems, extrusion uipuieiit and high demand for these products all ribuied to this spectacular growth. PVC resins produced in the U.S. span the inherent ity range lrnu O.lifj to Mb. Other resin properof iui|toi'Utiice in most applications are porosity, Wn rate, heat stability and bulk density. High volresistivity is needed in electrical application* and a 'tm gel mmem is required lor cerium flexible and rigid products. Kusiu properties required for different implications are given in Table 2. Three iyj>es of processes are used Idr PVC homopolpaer manufacture. These are NUs|>ensinn, emulsion iml mass (or hulk) processes. The proportion oi PVC cade by each process is now reasonably stable and ifcouid remain so in the next few years.'' About HO percent of all PVC is produced by suspension proasses. 'Phi* review is limited to u discussion of PVC Hupciisioii processes. Ife chemlttry of vinyl chloride potymcrlntton. Vi- syl chloride monomer (VCM) is- polymerized efli- oenily in die preseme of a feet* radical source. The putit'uJ reaction scheme and initial kinetics are typical free radical chain reaction- Initiation Initiator--K-*- 21 WO 1975 'MoOetn Plastics Annual summarm* Fig. t--Growth of lha PVC industry. Polymerization + M Jin* + mJ&IK' 1978 Chain transfer to monomer * R + M PVC 4 R Other chain tran.der agent (CTA) Termination R- + CTA--i-PVC + R 2R' --PVC Overall. |>olymeri/ubon rate is influenced Ijy the initiator (nr initiators) used. Vinyl chloride polymeriza tion is inhibited by oxygen, nitrous oxide and other free radical inhibitors. (Complex kinetic* arise Iteeuusc of dilTereni |Kilymcri/aiHm rate* in the mouomei' and ixdymer phase, Termination reactions <Kd are re tarded because of reduced radical mobility in the polymer phase- (amscquctitly, the overall rate gener ally increases with the amount of |x>)yiucr Idrnud (Trommsdorf cITcci) a* shown in Fig, l. Soiue com panies have developed initiator combinations to pro vide for more uniform reaction rate throughout poly merization.*-* Thu molcruhir weight of PVC dcovasc* with in creasing fxiiyisterizauun temperature, i.e., K., increases more rapidly than Ky. The relationship of polymerisa tion temperature to a measure of molecular weight {inherent viscosity) is shown in Fig. H. Polymer molecular weight is nearly independent <d tos CHI Rmh-- Swvc* Cvo 39 XP00010783 I TAELS 1--Listing of PVC producer capacity, third quarter 1979 (*Cpra*na nama ptatt capacity, MM Iba) NORTH Diamond Shamrock Aftarta Gas ESSO Qmrtat Canada I. F. GoodneO (2 locations) Total Industrial Rasistoi (2 locations] PiMticn Omaga Poheyfl (2 loeations) Potfmaras da Modes Promodonas industrial*! Itadcmii Total Air Products and Chantieate <2 locations) Bordan (2 locations] CartiMviMd Conoco Chamlals <2 locations) Oiamond Shamrock (2 locations) Ethyl Flrastont (2 locations) Gsnani Tin ar< RuMar (2 loeations) GaorgMPaeiflc I F. Goodrich (6 locations). Goodyear Tin snd Rubbar Gr**t Arnsriean Chamical Hookr Chamiciii and Putties (2 location*) International Materials Korsof-Century - Panaaote (2 loeations) RieoChimtais Shlntech Stautter Cnameai (3 locations) Tanneeo Chamicals (3 loeallont) union Cartiida Total COmULAMmCA Potycau INTO Etectrodor SACI indues Vlpiastie Total GraM fteaknt VMtea C.P.C. - Patioquimtca SAOmdolnsa NMM QUMMGti MUIU vWP Tom -7,200 $5 es 40 ___ 15 -120 100 300 90 165 650 Soctedad Pmmnp Plastioas Patroquimica WtSTBMt EUMPC m Beldam BadtohK Some Total Fraaaa Sectete ArtoaMiM da Vinyl CdFCMmlo AT0 CMffite Pradulta Chlmtouat Ugh* Kuhlmsnn Rhone Routine (3 locations) 8imh Chimte SoMc Total Waal Gomany BASF Charnteob* Works Huols tci Sohiy Hoechst (2 locations) Lena wackar * Chamla (2 locations) ToM Italy Socteta CMmlca Ravenna Uduktubniea Famndina Montedison (3 toeationi) Rumisnea Sud Soaata IMiana Raslne SoMc Total MtUartaada DSM snail UMtedCofdM BP (2 loertons) Britten industrial Pttstica tCI Vlnsts* TOW GTROt WBTBM fUUOPt HaMe Kunsttwffworka oy Pttroowlnitea - Dow CotombUna do Cardura y Oartvatios Patroquifftica CotombUna 30 ESSO Pappaa Chamical Hofttwam Gras Ctwmlcal Induatry 17 uwy 60 Norsk Hydra -2J -2.4 -2,' initiator lyjic but in somewhat dependent on initiator concentration.m However, in addition to the tempera ture dependence, molecular weight can he reduced by a chain transfer agent (CTA). A chain transfer agent reacts with the growing polymer chains (R-) more rapidly than monomer (K, >>>K,) terminating that chain and initiating a new chain. Common chain trans- AO fer itgents are thioglycolatcs11, halogenated hydnxanl bons'* and aliphatic aldehydes.13 ] Initiators used in suspension PVC processes are* generally peresters, peroxydicarbonates, peroxides nr ^ a7.o compounds. Initiators arc chosen to have decora* | position rates that produce polymerization rates to match reactor heat removal capabilities. Other comkt* MARCH I960 HYDROCARBON PROCESSING APobbiom ir 720 to 373 ^55 300] 09] Oi\ 230. 700 300 p] -2,520] 39j MO! 2 450 600 700 -175 300 Its 1.300 230 300 220 -175, 330' 400 575' 200 - 750 125 1.650 130 110 100 130 16S Companftit inoustrSI 4t Mw SNMeas Mia Enargia y industrial Aragonaaaa Hwvfc Industrial 12 locations) Monsanto tiMfia (2 locations) Rto Rodino (2 locations) Total Kenwwrti (2 ideations) tmuuai Iona UflSW EUROPE Stata Industry (2 locations) Stata todiMiy (2 locations) tatf (may VEB Chumlseht Worts StM Industry (2 motions} load Slat* industry (4 locations) SUM Industry (4 locations) mi State Industry (S locutions) THilUdi Homijrta tnd. (2 locations) Jugovtnll Oreutko Homijski tnd. PoJychum Vinyl Piastika Otu UtT Electrocftamkil Industries Tartay town (2 locations) MUST Ma Ahmsdabad Mamrfacturtng ChsmioN and Plastics India N0CJI ntnci wm tc cnwHcvi toasthan Vlrryi P. T. Esstam Potymar p. T. Sttndvd Toyo Polymar Autti Glass Central Chemical CMsso (2 locations) CWssd Ntrochanticit 50 300 65 350 230 220 450 $00 500 1.100 175 ss 25 125 45 200 33S 45 45 45 30 60 35 50 50 50 200 100 DENKA (3 motions) KanogafKM Cbsmicai (3 motions} KiwimM Organic Chemicals Kuraha Ctamleai Mitsubishi Monsanto Mitsui Tostsu Nippon Zion Nissan Petrochemicals Rye NkW Shirt - Etsu Chemleal (2 motions) Sumitomo Chemical (2 locations} Sun Arrow Chemical Toagosai Chemical Totaiyams StUsui Toyo Soda Total Marti Kara SUN Industry Mk KarM KPIC (6 motions) Lucky Malayan Malayan Bedo Chemical Synthetic Rosin POItM Araksy Phllipoiaat Matiuhay Vinyl Phitipptna Vinyl Consottiutn tlMsyira Singapore Potymar Tatwo Cathay Plastiea cnina Gmf nasties Formosa Plastics Ocean Plastics TtaUund Thai Plastic and Chemical OCEANIA Aastnlta B. F. Goodrich ICt AFRICA AMarts Sonatrach Ubpa Hatmnal Organtzation tor IndwrtaHatton SNEP juries AE A Cl (2 motions) AE 4 Cl and Santracham 1- wan n HH CeawwaiM Iwan M ... Wwu Mpimb Inci 375 450 50 265 240 260 42S 50 250 480 220 130 100 100 200 -4,000 50 360 100 25 25 10 65 4$ 35 SO 80 650 40 40 135 230 60 130 55 100 220 rnc.ir* s are tie* or ecomte* to onsid- :ssino nations are effect on polymer heat stability, safety and convenience of use. Suspension polymerization requires that vinyi chlo ride droplets of I00-I50p average diameter be dis persed in a water-continuous phase by use of suspend ing agents and suitable agitation. Typical suspending agents (often called protective colloids) for use in PVC manufacture are water-soluble polymers such as 21 modified cellulose or a partially hydrolyzed polyvinyl acetate. The suspending agent concentrates in the vinyl chloride-water interface and prevents (or minimizes) droplet coalesence. The nature and quantity of sus pending agent and the agitation employed are major HVMOCAMON PNOeCMMQ MARCH 1MO 41 APOOOI0785 / TABLE 2--^VC applications--Ruin raquiramanta RESIN REQlHRCNENTt mam imuatiim* viemity Nrtll RfHUm !. Extrusion Rigpidiu(tntoonM* ) Pp* Profile Sneet 098*0.95 O.06-U2 0.74-0.96 (iln 0.74*0.92 RottM (PtttMUM) Wire anc Cette i.oo*i.i2 Prolim Siwet Film 2. IfyKturi MoMtng Rigid ftaeM a. Blow Molding Genies 0.90*1.12 0.90-1.04 0.90*0.91 0.60*0.92 0.90-1.04 0.60-0.92 4. Calendering fteutte Steel 0.72-1.04 Hign Hut S(ability Hign Hut Stability Hign Hoi SuMrty. Hign Clergy. Good Early Color, low Gets Hign Haai SuMtty. Han Clarity. Good Early Color, low Gats. Sort Electncil Proper- not. low Gaia, Hign PlaonoMr low God. AHbanornanaoariiie* 2r Absorption. Uw Gad. Higtt Cdnty, Good iariy cotor Low Gad. High Canty, Good Early Color. Htgh Hail SUMity Hign PlasMuar absorp tion Hign Haai StaMity. High Clarity. Good Early Color. Low Gad Hign PitsiKuar Abtorplion. Low Gad. Good Early Color (actor* in determining resin properties such as particle size, particle size distribution, porosity and bulk den* illy.* Kliccts of agitation in the su>|K,u.sim |x>iyiiicii/atim <l vinyl chloride liave recently been correlated with resin properties by use of Weber number.14 Other variables of importance in suspension |x>iy* liicvi/aiinn <>l vinyl t hloririt* are iinimriiies *j>icwnt, imliT ni addition of rciuiioii maierials and rale <l mixing a> well as laic of heating to reaction tempera* ture. A typical procedure would call for removal of oxygen from the reactor by evacuation, addition of water, vinyl chloride ami suspending agent and finally addition of initiator. Polymerization begins as die sys tem is brought to reaction temperature. Consistent luuchno-hatch results in PVO manufac ture require control of essenuaily all variables. These include impurities (oxygen, butadiene, etc.), rates and temperatures of additions, thoroughness of mixing H.ilk rU'ittn, .umI pmMlr ,rr aka, ilr|itmhi <*. ihr rxiinii ,4 II*W|`*,*1 Successful PVC manufacture depends upon closed systems with reactor internals being cleaned without opening any vessels and complete removal of any remaining VCM. PolymsfizatlOA temperature, 'F Ftf. V-PVC inherent viscosity versus potymenudon tampersturs. and rate of heating to reaction temjveraiure. A com plete description of a polymerization procedure in* eluding polymer isolation is given in a subsequent section. Interactions between polymerizing monomer drop lots and also inicraciionx Ijctweuu growing primary polymer particles lead to differences in the structure of PVC. PVC morphology is affected by the suspension system, agitation, polymerization temperature, nonreactive additives, conversion and other factors. Differ* cnees in PVC morphology are illustrated by the photoiuicrugraplix in Fig. 4. A uuiiortuly porous resin (4Af is desired for ease of VCM removal and processing. Large gel regions, shown in 4B will release VCM relatively slowly making llie* shipping pit*ess less elleuive and may cause non-umlormily in the final product. Commercial proceaeaa. Polyvinylchloride plants in the 50s and 60s contained many small lunch reactors grouped in one or more reactor modules. A typical * MAIICM 1W0 HYDROCARBON PROCESSING AP00010786 r i i A. Uniformly porous room .4--Seaming Moron micrographs o PVC ratio croaa taction (500X). B. An ooMSmng s latge gel reactor module conutined twenty 2.S0()-g;tll<m rs. PulyniLM i/uiiiuii began with addition of water, ending agent and initiator to an oj>cii reaiitor at acai Londilioiih. The reut-ior was then dosed, evacI to remove oxygen, vinyl chloride added and the tture agitated and healed to |ajlymei'i/aiion tent* ziuve. (lotistam tetn|x.`rature was tiiainiainerl until WU)HT4vnt o| ilie monomer had been converted to tlytner. Unreacted vinyl chloride was removed by ration and the {Ktlymcr recovered front tin- water by tenitilugiug and drying. Reactors were etted and vlvanetl manually after each cycle. Tlte cycle took approximately 12 hours. [ frucesKCs using larger reactors1*1'11'17 were dev<>lo|x'd llit1 late (MK. Hu- i.tigci re.Ktors nlicr increasetl Jut'iiviiy, lower plant costs and improved product liiormiiy. Although continuous PVC processes have en patented,none have been commercialized to With die discovery o! the link between VCM ex|x>and angiosarcoma of the liver, OSHA quickly ' regulations in 1974 limiting employe exposure, uerrs implemented many prexexs reiiueincnts iKtrily aimed at reducing or containing VCM entisKeactor cleaning emission* were reduced by loping waierdancing equipment that would operin a closed reactor and chemical treatment to :e rcacior fouling. lu the fc.l'A declared VCM to be a hazardous pollutant and issued an emission standard in 1976, g the VCM concentration in the resin slurry, .or o|>ening emissions and other plum VCM ends* in. Pnxim ers liegau developing stripping processes Wer the VCM emissions and fugitive losses. Rapid lylchloride removal by steam stripping required that resin be highly porous.** Therefore, dense nonuuk rosins previously popular for rigid applications |n been phased out in favor of porous, more easily Ji|>ped resins. f)o control VCM exposure in downstream laUritadtops, the industry established a 10 ppm VCM TOOCAABON PROCESSING MARCH 1M0 HISTORICAL Although polyvinylchloride (PVC) wu first made in the laboratory about 100 years ago, commercial production of the homopolymer did nut beKin until the late 1930s.' A major reason Tor the king delay between discovery ami MMiMucrtiab Malign was that unmodified PVC is not a useful material. Progress in PVC technology has involved s complex interac tion between resin development, effects of additives ami equipment design as well as market demands. In die past decade, a major new requirement has appeared--that of meeting PA and OiiHA regulations. To appreciate the latest developments in PVC technology. it is useful to luirdy review {M*i inilcsumes in PVC Fuundaiiunsofthe modern PVC industry were established by a series of discoveries extending back about 50 yean. 1, Initial commercialization of PVC homopoiymer was made possible by discovery of the plasticizing effect produced by esters, such as tritoiyt plwasphatc or pluludsites.* World War 11 niiniulaicd interest in replacement of rubber by plasticised PVC. Use of plasticized PVC as wire and cable insulation was common within the decade as was use of calendered PVC as a leather replacement. By 1957, the fledgling PVC industry had tales ofover 64)0 MM pounds,1 mostly in flexible applications. lathe early I960*, advances in heat stabilizers and extruder design contributed to the rapid growth of rigid PVC. Growth of rigid PVC (pipe, fitting!, construction profiles) from 1965 to the present has been spectacular (Fig. 1). The discovery in 1973 that exposure to vinyl chloride resulted in increased incidence of a form of liver cancer (angiosarcoma) resulted in fundamental changes in the PVC industry. A result or compliance with EPA and OSHA vinyl chloride emissions standards haa been loss of capacity, cnimated to be alioui 90 percent4 which tomributixl to PVC shortages in 1976 and 1979. However, technology to control emissions now has been developed, and a new round of plant expansions has been announced.1 In 1979, U.S. companies will have about 7 billion lbs of PVC capacity (Table I) with world capacity about 30 billion lbs. Supporting the rapid rise in PVC production has been improving technology for vinyl chloride production. This has recently been reviewed and will not be discussed here.* 43 <t i AP6o6l6787 / RMCtor design is important in rosin quality control, along with agitation, speed of agitator, baffle location and ability to move baffles from outside the reactor. concentration limit in dried FVC resin and arranged for surveying and enforcement by the Plastic Pipe Institute. Typical wcpcncton PVCprooccc, A PVC process can be divided into polymerization, monomer stripping and resin drying sections. The process flow of a typical large reactor plant is outlined in Fig. ft. The process begins with the charging of process water, fresh VCM, recycle VCM, and suspending agent to an air-free reactor. Charge quantities of each must lie accurately controlled to avoid changes in the resin panicle sire and other quality parameters./!"he charge is controlled by How measurements, lews of weight measurements, or a combination of both. Water to monomer ratios of 1.1 to 2.0 are normally used depending on the quality of the local water and the polymerization technology used. Monomer soluble initiators) is added to start the reaction. The heat of reaction is removed through a water jacket around the reactor. Initially, hot water is circulated through the .jacket to help bring the re;* contents to the desired polymerization irm|*ir;ii Cooling water is then used to control the rcn> temperature. As the reaction proceeds, a stable suspension VCM in water is formed by agitation of water, V and the suspending agent. The product panicle and distribution is generally Ixdievcd to !>e extahlix early in the reaction. At a conversion of around 70 percent, reactor p sure (and reaction rate) begins to drop due to reacted monomer being absorbed !>v the- PVC p.-ini (Fig. 2). I lie reaction is usually tcntmiau.xl at a spc> conversion measured as a specific drop from re* run pressure or at a final reactor pressure. The t monomer to pojymer conversion ranges from 7' percent, depending on product quality requireme Unreacted monomer is recovered from the I product by depressurizing the reactor to a reco' system containing slurry disenirainmenr equipm vacuum pumps <uid compressors, flic vacuum pui and compressors are usually water-sealed. The reactor is rinsed with water before rechart to remove residual resin which, ii left in the rent would degrade die quality of the resin produced in next batch. Hydroblasting may be used to remove * that has formed. The reactor surfaces also ma> rinsed or sprayed with chemicals which help pre polymer scale formation on the reactor internals. F.1*A standards sficcify dial die residual mono content of the resin be reduced to 401) ppmw m mum by additional monomer recovery in a strip] process. There the slurry is heated ami sirippet contact with steam in a trayed tower. The stript pvc reactor recovxreo MONQMCR STORAGE BLEND TANK STRIPPER FLUID BED ORVER S0.O 44 MARCH taao HYDROCARBON PROCES AP00010788 icior re. ctor n of /CM sir* shed ores* un* tides rcific ictor final 3-w ent*. PVC very teni. mps KnK ctor. i the scale Y be vent mer taxiping I bv process must be designed to minimize ihe heat expo sure of the resin to avoid heat stability (resin degrada tion) problems in customer processing equipment. In the drying section, the polymer slurry is centri fuged to produce a wetcake having a water content of 18-25 percent wt. water. The range is due to variation in die intra-particle resin pore volume. A two-stage fluid bed is used to dry the resin to below 0.3 percent *t water content, 'rhe firm stage is tackmixed and operates in the constant rate-drying section of the drying curve (Fig. 6). Approximately 90 percent of wetcake water content is removed there hy the fluidiza tion air. Water at MW-220*F is circulated through heal exchangers in the hackmixed section of the fluid bed to provide HO percent of the evaporative heat require ment with the fluidization air, at a comparable temper ature, providing the remainder. The second stage of the dryer operates in the failingrate region of the drying curve and is designed to approach plug flow of the resin. Heated fluidization sir, again between 190-220*F, provides the heat input. An air cooling section may be added if the resin is particularly sensitive to heal. Normally, there is little, if any, degradation in the heat stability of a resin across a fluid bed dryer even though the resinTesidence time is 50*120 minutes. Evaporative cooling balances the heat transfer in the tacktnixvd sertion >r tlx* dryer. Rosin temperature does not approach the drying air temper ature until the outlet of the plug flow section. The product resin is usually shipped by rail in hopper cars, but some is shipped in bulk by truck or in hags or gaylonh. I he rail cars arc epoxy-dud to prevent resin contamination with rust. Ail pan* of the process contacting the resin are either stainless sieel, aluminum or cpoxy*clad carbon steel for this reason. Some processes use glasslined reactors. IWirtOC dMiQn. Reactor design is governed hy resin quality :irnl e;unir ptodiutivily ivqoirciuctils ;md plant production considerations. The polymerization rvci|x- determines many ;isjx.fLt of resin quality. However, recipe changes cannot al ways correct for deficiencies in reactor design. Agita tion is the primary design varialrJc affecting product quality. The mixing blades are usually either retreat curve or turbine types that pump the slurry from the bottom of the reactor upward# around the sides of the reactor. Reactors are usually taffied. Agitators may lie tingle or variable-speed and the reactor baffles may be externally adjustable. Quality can also be affected by reactor size and configuration. The variation in product quality from hatch to tatch can be reduced by using large reactors <ince it is easier to control several large reactors chan many small reactors. However, it is necessary that the process be well controlled since many more (toundx of off-grade resin are produced by a bad butch in a large reactor than in a small reactor. Productivity is maximized by initiator choice and reactor heat removal but may be limited hy quality considerations. Typically, reaction cycle times range between seven and ten hours. Of this, reaction requires `0-80 percent with charging, degassing, dumping and turnaround taking the remainder. Obviously, the larg est potential for improved productivity is in imitirni/iug |K>lyiiicri%iiiion lime. Polymerization time is limited by the reactor's heat removal capabilities and bv the polymerimtion hear release curve (Fig. 2>- HifltMcm initiators cun Iw dtosen to produce a beat release curve that best matches the reactor's heat removal capabili ties. The heat released during |to1ym<*ri/ulioM is <omimonly removed from the reactor by circulating chilled water through a reactor jacket. .HperiaHv-destgiu'd cooling taffies may also be used. The slurry side jacket heat transfer coefficient decreases rapidly as the monomer conversion approaches tit) jx-n cut (Pig. 7). The decrease is dependent on the slurry concentra tion. Since the wall thickness of large reactors ap proaches one inch, the thermal conductivity of the reactor wall is an important consideration. In ibis regard stainless steel-lined carbon steel has an advan tage over a solid stainless steel or glasslined reactor."-" A few producers have developed reflux condenser technology to improve rhe reactor's heat removal capa bility. VCM vajiorized from the reacting slurry is condensed in a vertical condenser mounted externally to the reactor. Condensed VCM is returned to the reacting slurry. Since the condensing (teat transfer coefficient may Ire higher than the jacket coef ficient and the condenser area is not limited by reactor di mensions. plant cooling water can be used instead of chilled water. Initiators that give nearly uniform heat release throughout the reaction are used since reaction heat removal is not heavily dependent on slurrv jacket transfer coefficient. Reactors with condensers mav have productivities of 2,500 Ibs/yr/gal reactor volume StNO KVOAOCAftSOM PAOCfMIMO MARCH 1MQ 45 4 i i i i i i i AP00010789 / With the trend to larger reactors, there has been a and oilier HupjMirt systems. I In- snp|iori svmchu .tr? used more efficiently when they approach communm 1 usage without bottlenecking the reactors. 1 corresponding trend to remote Stripping options. The process selection lor stripping 1 residual VOM (KV(!M) (mm the resin slurry is depm* A or automatic valves dent on the resin characteristics, quality requirement*, B and the final R.VCM concentration required. The resin S with many safety Interlocks. VCM/PVC equilibrium relationship is important in B stripping. Ileal .stability sperilVaiMm* limit die lira B I history of the resin and may limit the number nf w options that can l>e considered. ft or more. Reactors without condensers usually have In evety case, the resin stripping de|iend* on a productivities under Ihs/yr/jpd. combination of time. U`iii|K't;utm< and purge gas (wa* Plain piodnclion considerations can also ulltri rear- ter on nitrogen) to remove VCM from the rosin. Mn*t V lor design. To minimize plant capital investment. die pnaiesses use stesim to provide the heat input and V reactor si/e and tminlier of reactors in a module remove VCM. Processors have developed new recipn 1 should he balanced with the sire of the recovery system and additives to make resins easier to strip and in * prevent resin degradation. I I Of the stripping ojMions in use, the most rornnHin are Ixiith and continuous stripping ol slurry. Steam is 1 injected direedy into the slurry. In batch stripping the slurry tem|>crattirc profile is adjusted to control the a RVCM of the pnxluct slurry. Continuous stripping m jnay be done in a trayed-stripptng tower, again using steam to heat and strip the slurry. The continuous a process may have advantages in providing more am* I sisfcntly low prodiui RVCM. Also, plum production can lx: increased by eliminating an in*reactor stripping time requirement. Both batch and continuous strip I ping are generally conducted in a wav to minimize cxpistire of the slurry to high lempcnitmrs. With a V continuous process, this can !>e done by beat exchang* ing against the feed slurry. B Oilier scheme* using inert ga* snipping ol die slum 1 or ol lilt* centrifuge wetrake have been proposed and * ttiny Ik; iti use- 4 qI- i i 0 20 40 i . i. . i. _ii__l 60 80 100 S Conversion W* 7--Jacket inaid* heat ttanatar coefficient v*r*u* oonvaraion. Fig. --Typical rotary dryer arrangement. Drying options. In addition to the two-stage Ihiid Ixtl 1 dryer discussed tit the typical process section, rotarv B and flash-fluid bed dryers are used, The type of dryer | selected depends upon the dryer si/e, product pro|)cr* | ties, quality requirements and fuel availability mul I costs. 'I Rotary <lryer (Fig. R) utilize ci>-ciirrent flow or , wetcake and the drying air. The dryers are designed J with internal-lifting flights and baffles ro promote contact Ix'iween the resin and the drying air. Thermal ' efficiency of a rotary dryer can approach that of a Him! : bed but only when high temperature drying air is used, 'l'he air heating can be done either tty direct-fired natural gas in an inlet hunter or l>v iudireci means through beat exchange with steam or hoi oil. Direct oil , firing is questionable Irecaiise combustion products might downgrade product quality. Rotary dryers can have initial cost advantage in the smaller sizes. 'Flie disadvantages of the rotary dryer are dial its high inlet tetiqKfratiire can burn resin and H* thermal eMVieiirv is lower than that of a fluid lx-d dryer. A nash-nttid bed dryer (Fig. U> is ;iwo-*iagc drver that uses a high lemjx-nmue air si t eam m cut rain tin* wetcake in a duct and dry it through the constant rate section of the drying curve. The resin is recovered in a cyclone and the drying completed in a plug-flow 44 MARCH 19*0 HYDROCARBON PROCESSING ^ AP6b6i0796 FLASH DRYER FLUlO BEO DRYER BAG FILTER \ * fluidized bed. While the fluid bed section should produce a more uniformly dried product than a rotary dryer, the Hash section has the same potential resin burning disadvantage as the rotary dryer. Dejxrnding on the drying air temperatures, the thermal efficiency of the flash-fluid bed dryer should be comparable to or better than a rotary dryer. The two-stage fluid bed dryer has the advantage of never exposing the resin to high temperatures. Also its thermal efficiency is the highest of the three type* of dryers and its heat input can lx.1 provided by steam. Since its higher thermal efficiency results in higher humidity exhaust air, care must be taken with dryer insulation, heat tracing and in the dryer operation ro prevent condensation in the exhaust cyclones or iugitou-ses. Safety Itpactl. There arc several aspects of a l*V(l plant and its operation that merit special attention from a safety standpoint. First, being a hatch process, there are many reactor startup and shutdowns occur ring everyday. This increases the risk of a mistake that could endanger lives and equipment. With the trend toward larger reactors, there has been a corresponding trend to remote or automatic-operated valves. Pro ducers have added many safely interlocks to thrsc automatic valves to prevent them from lx.*tiig opened nr closed at the wrong part of the hatch cycle. Also, provision rnuM lx: made lor controlling the polymerization in an emergency situation such as a power failure. All prHlurcrs have chemical injection wstems that will chemically stop the polymeri/nthm in an emergency. Some prodmers provide diesel or ueattt turbine-driven cooling water pumps to provide lor reactor cooling during a power failure. T hose having reactor condensers can benefit from cooling even without agitation. Some producers elect to pro vide back-up power for the reactor agitators and the cooling system, The important point is that the plant must have a reliable and effective system to control' polymerization in an emergency. Of course, reactor pressure relief systems must be designed to control pressure in the event that a run away reaction did occur. The design oF the relief valve discharge piping must carefully consider the flowing characteristics of the different multiphase flows that could be relieved at different points in the polymeriza tion. Several articles1-*-*5 discuss reaefnr relief svsiein requirements. Initiators used in PVC manufacture are generally unstable at room temperature and must be stored at low temperature*. Suppliers have s|K-<ili< itiominendaiiotia in this area.*'-2' Operating procedures are important in insuring that the initiator is kepi relri. gerated when being transported from the storage freezer to their process area. The recovery and initiator injection systems must l>e designed to prevent initiator carryover into the recov. ered monomer system or initiator injection into a monomer line or vessel. Vinyl polyperoxide* can sometimes lie found in the recovered monomer system. .Since these compounds can self-detonate under certain conditions, care must be taken to avoid concentration when clearing equipnicnl of VCM. Recovered monomer can ixr c.iuslictreated to destroy polyperoxirles. J'VC dust can lie ignited by a very high energy discharge. It becomes more flammable if mixed with mmlmKtihlc additives, such as plasticizers. because VCM is a Hammalde gas. the reactor area in mmt plants is equipped with u water deluge system designed to help dissipate a VCM vapor cloud and to cool equipment in a Are. The deluge system is auto matically activated by combustible hydrocarbon ana lyzers and by heat actuated devices. The VCM storage tanks are also protected by a water-deluge system. Reference `28 provides detailed recommendations in hvorocarron PROCESSING MARCH IBM 4? AP0001079I f \ 1 50 nm > t r rr t reduce exposure to or l>elow the permissible level. Thr II 30 20 standard does not specify what controls are to lie used. Producer* have altamcd compliance with tin- OSHA standard by modifying equipment clearing and open*ins procedures and by eliminating VCM emissions or s *t 2 10 collecting them in vent headers. Ventilation of build* ings ana equipment is also used. !, The KPA VCM standard limits the average VCM content of the stripped resin at 4UU ppmw and the. . J --1--1_Li.i.1l i I ,.l i 1j_ reactor opening loss at 20 lbs per million lbs of resin 1 5 tO 20 30 SO tOO 200 produced. Other equipment opening losses are aim Cumulative production tuition pounds Fig. 1^-Value added manufacturing experience curve for PVC in the U.SA. 1953-1976. controlled. VCM emissions from vein and water streams it|>*lrcaiii tif the sttipping step are limited mt concentration of Id ppm. ( lie standard also requites these and a number of other safety areas in PVC plants. that there be no manual or relief device discharges to the atmosphere from equipment in VCM service with the exception of an emergency relief. VCM regulations. in wklition to air emission ami water discharge regulations in general, PVC producers must also comply with the OSHA-employe VCM-expostirc standard**1 anrl ilc KI'A VCM-emisskm stand ard.TM The OSHA standard established a permissible em ploye VCM-exposure limit of 0.5 ppm and a maximum exposure of 1.0 ppm over un eight-hour periml with out regard to ihr use ofresjHnilor*. htnph>ye-cx|KMurt monitoring, training, and medical surveillance must be provided. Also, the producer is required to implement feasible engineering and administrative corfrml* m As previously discussed. VCM-stripping procedure and equipment have Men installed to control the VCM content of the resin slurry, and purging procedures used or process changes made to tomply with the reactor oftening loss limit. The VCM comcm of water streams controlled by steam stripping. Incinerators, carbon adsorption or solvent absorption systems are used to control vent emissions. Tlx* KI'A standard differs (root the OSIIA standard itt that it requires specific controls lie used to control fugitive emissions. Pumps and compressors in VCM service must have double mechanical seat* or he seal less. Agitators must also have double mechanical seals. Rupture (lists ititiHl Ik* ittstalled untlei VCM relief TASUE 3--Polyvinyl eMorid* price and production history IMS 171 valves that discharge to the atmosphere. Equivalent controls may lx* iiM*d will) KPA approval. flic iiKlustry has louixl that the cost of compliance with the KI'A standard, measured Ixtih in capital utxi #.*. * fill laiMaa Ywr Mi) Mwr ^ VMM (1972 tint 4/M) tm FVCprtM (1972 (tWTMt *> WM> in lost production, is significantly greater than that of the OSHA standard. Manufacturing efficiency. Mg. I o shows value added Prior U.k. 3.000 ha HA NA NA manufacturing experience curve for PVC production 1953 43S 3.435 NA HA 65.9 316 S1954 1955 397 3.932 4.359 NA NX 17.0 39.9 63.0 37.6 65.1 34.0 1959 4.911 172 31.4 <66 30.6 in the U.R.A. between lt)SS and 1!>7H. This curve expresses a relationship Itetweeu cumulative PVC |>niduction (experience) and the "value added," or PVC 1997 999 5.496 16.9 29.1 44.9 29.2 1959 6.342 1M 23.3 40.1 26.5 price less the VCM price, with lx*th adjusted (o a i S81i 1969 1999 1911 7247 6.1*3 9.K0 19.1 20.9 14.S 19.0 11.7 14.2 37.0 25.0 33.6 23.2 27.9 t9.3 constant dollar basis. The entries art* plotted on tmih rnorriiiiaU- axes as logariihtoK values, litis graph, IS 1.211 10,373 1914 1:8 11.791 13.366 199S 1J37 15,235 1999 2AI4 17.396 10.9 14.3 1.4 15.5 9.7 15.4 12 15.1 7.7 13.9 24.9 17 6 25.3 11.1 24.1 17.5 23.3 17.3 21.6 16.6 prepared by Townsend el aP* plots historical price and production information (listed in Table 5) gathered by the U.S. International Trade Commission. The ob* served yearly prices iire adjusted lo a 11172 Ixtsis by the 1997 1999 1999 1970 1971 1979 1973 1974 1975 1979 2.149 2.935 3.032 3.115 3.437 4.322 4.594 4,744 3.965 4.545 19.541 22.179 25.209 29.323 31.760 36.062 40,676 45.420 49.395 53.936 9,7 13.2 5.6 11.4 5 1 10.7 4.3 10.5 4.4 9.9 4.0 9.0 4.0 10.2 7.1 12.7 7.9 11.1 7.5 11.6 11.9 15.7 17.0 14.0 15.6 13.7 14.6 13.5 14.0 13.4 13.0 13.0 14.2 16.0 19.6 23.0 16.9 24.1 19.3 25.9 CNP deflator. `The "value added" over and aitovc the raw material cost in constant dollars is a measure of the efficiency nT the manufacturing step. The raw material >x(H*iise k taken as the cost of l.Olbsol V(;M |H*r jiouiidol PVC resin pr<Kluced. As tnunufarumiig ex])cnse de creased with experience, as for example by increasing 1977 5.267 59,147 1979 5.B7I 6S.075 1979 (Mt) 6.050 9.7 10.0 94 6.6 16.7 26.5 17,0 256 the scale of operation* to dilute fixed costs per |XHmd of output. the "value added'* declines. The experience curve covers a 25-year historv of halMWIWM tHMHCtf W HWMI CBIM It I k l>V(l|l 0MBMtMM0MIH'SMatf PVC production, and the steadily declining value added is indicative of' the improvements made in the 41 MARCH 1M0 HYDROCARBON PROCESSING AP00010792 f 4 e manufacturing technique* during this period. The J. discontinuity occurring between 1974 and 1976 is 1A believed to he due to increased energy emu miri it- fovcrnmoiu rcgulaiion*. or 11 prcKlucers are to continue increasing the effi d- ciency of pvc manufacturing, further process im provements will be necessary. It is likely that contribuM linns in this area will include: (a) replac ement of small ie inefficient j>nx.esi plants as they laecome obsolete by tn hrger scale equipment to reduce fixed costs; iO lb) fun her improvements in raw material and energy rr efficiencies; and (<) more productive equipment utiliz ing faster rcaiiiom, improved process rontrol and more sophisticated means of dissipating the heat re- leased by polymerization. h Manufacturing coat Table A gives an illustration of the cost structure lor suspension HVC manufacture. s The cost shown is for a hypothetical 200 MM IbVyr i name plate (175 MM actual) plant beginning operation s' in the conditions existing during the third quarter of 1979. The plant is assumed to tie located in the U.S. r Cull Coast area. 'Hie plant capital cost is estimated Ss `>0 million dollars. Several companies who license their (uspension I*VC technology provided rum-confidential information hImmii their process Indore this artide was prepared.1* 'Hu* raw material and energy use and niau|nwcr requirement*, although ikh |4aUented after anyone's specific process, represent an approximate average of the information supplied by the licensors. The hypothetical plant is assumed to produce a typical suspension resin slate. This size plant would utilise four reactors about 22,000 gallons each in capacity, use re*m stripping for monomer recovery ami include recycle <>l unrcaried vinyl monomer. The plant would TABLE 4--Typical suspension PVC manufacturing expense new 300 UM Wyr targe raector ptert Gulf Ceeet V.SJL Third Quarter 1ST* conditions Praducr PraetM: l_ 2VC____ Wes swpsnslw in low fl.ODO-otete rwtten Staani itrtppim and racycta ot racowed fwiv miiiWt CafMAy vWzWok Actual capacity: CapM Inwlmart: 2Q0MMlte 97.5 pane* 179 MM Ite S90MM OaR MM tMMI Can mu a/ttpvc 1. Vtnyt eMoctda 2. SyproduMs 17.2 CM) 19.0 0* ICS (0.019) 30.093 (ate) 17 (0 23) SubtoW 30.490 17.40 1. CmtyRi Md ctamtaMi 2. ftontMOlM 3. Oaacauna tupodM SubloM UINNIM t. Paewoty 2. Staatr 3. Preens and ccoling water 2.09 0KWH 5.00 &Mt 0.17 1.2 Subtotal VARIABLE MANUFACTURING COST AT 97.5% OF CAPACITY Hourly later and i Mimtananea matamis and MPOtes (2.9% ot oottai) Oparatmd SuppUM ^ 1.090 07$ 17$ MOO 788 1,050 0.00 0.90 QIC 1.20 0.49 0.60 34,747 1.800 1.254 50 19.09 0 91 meet ail applicable environmental and worker protec SuttoW tion regulations. 'Die overheads, sales ex|x*n.sr and miscellaneous nms are typical ol published figures and are probably fairly realistic for an operation of this size. Total plum employment would l>c approximately 100, witn 75 Hourly and 25 salaried employes. The staffing level tew ot direct costs imartit on wertdrw caoRal (12% of SS MM) Rant owrtesds |if)0% ot later easts) TwteyM ****>d 4% ot cm*) Oapradaten (10% at capiui) SteMtll 000 1.000 790 _5.000 ~V.660 034 0.01 0.43 2.06 4.94 ctnild pmlxihly Iw reduced by approximately 10 per cent if the plant were to be located as pan of a large established manufacturing complex. Cor it pent gata SMn cams. M0 and narat Mwwsitnuvi (7.5% eUMs grwa) 49.997 4.100 20.00 2.34 'Hie computed product valve for PVO mnniifuclurcrl in this plant with third quarter 1979 conditions is 31.26 cents per pound, which is approximately the PmM and toeem tas (tp% at Haad emrtai) Profloet mm at am otus io% roi 5.000 94.007 206 31.20 market price at the time. Thi HVC price includes a pretax return on the investors capital of 10 percent which is relatively low. The returns on new I'VC capacity must lie adequate to encourage investors to share of total sup)>ly, margins will increase u> rellcct tlte higher capital burden of the industry as a whole. install new production facilities if a capacity squeeze is Future trends. PVC technology Has evolved r a rapid to be avoided.*1 Hopefully, several factors will help to pace during the past decade. An indication ol ibe improve rates of return on investment lor new facil dynamic ivaunc of this technology ys given by the ities (I) huger yale plants cat) Ik.* installed: (2) com number of announced technology improvements, panies already producing a large-base load !*VC vol plant expansions and license agreements. It is often ume can incur less burdensome fixed costs in their new difficult to determine the exact nature of the trcluiol- facilities as a result of further distribution of over ogy change reported because details are carefully heads, sales and management expense, and the like: guarded. However, the impression of a dynamic indus (3) continuing improvements in techniques will lead to try created by this activity, we believe, is accurate. cost savings in spite of the general inflationary bias: As production facilities at existing installations which and f4) as newer facilities become a more prominent utilize numerous small-sized reactors become obsolete | HYDROCARBON PROCESSING MARCH 19M 49 I ii i ! AP00010793 T ! and in need of replacement, they will lie replaced bv a small numlier of much larger reactors. The capital emi associated with these new reactors will he partially 0IIV1 ly the saving* in Hardens! dial die large reactors provide. Gains in product consistency and quality, improvements in controllability and improved plant safety will be side effects of more widespread use of large-scale reactors. i lte evolution to even larger reactors wiU encounter mechanical and logistical constraints. An optimum range in which benefits of Urge reactor technology are obtained hut operational problems are avoided will he csutMisheri. Suspension and initiator systems will Ik* improved to achieve optimum results from large reactors. This will allow most grades of resin to lie made by large reactor technology. Clean wall systems, VCM containment procedures anti emergency kill systems will continue to Ik- improved. Improvements in efficiency from the best presently available large reactor technology over the next decade will prrriutbly lx* modest. I Jowrvrr. in the highly competitive l'V<; industry, ;Klvuniagc* pre sented by advanced large reactor technology over small reactors or first generation large reactor technol ogy will Itc significant. PVC Iktntort. A listing of companies who have li censed some phase of their PVC production tech niques is shown in Table 5. The list is probably not complete us it merely represents those companies o About the authors Jack |t. CAMKHt.xia t* pmjrri Hiifirn'iiair, ('on- urn I'hnniniln, IInn*fun. M* ilutirn lurhulr fintrruK nwtlmiiuiH anil minmnirn of culn rrHiniifttflHrV. (Jamtruv received hi* ff.S. itt rhtmifttl rrtjintiring from fhr Hnirrmitff iif hfinmmri, ftrfmr jfH'iriN(| foMitm fir iwa with IMP. Ai.i.ah 4. l.iNiiki.N i* thnrhtr. Phulir* Hr.nmrrh, ('imnrti fw., Ptmtn d'i/jf, ftkin. fir. Mm Aru worked in rjptumlnrg ronmrrh iutil nr n mmrrh thtmiat fttr Cnttnm. Ine. Hr rrrrtmf him Ph. H. In nrynnir rhrntintrg from Kirr. Ifninmnlg ntuf hi* it.$. in rhrrnmfri/nml mntknmatienfmm ,twdwtrrw fWItv/r .I.kc H. M'Ot.i>.Y .]k i Htijierrinintf fmier** rrMfinrrr frith CttHorn. fur., Pnnf Cihj. 0iln. Hr in rr*tnin*ih/r fur *nfien>i*iiii of liir PV<` pnirtt.a ilfMiifini nmt /ihml Afr. Mri'nllrg rrrnirrfi Inn H.S. in flirniicril mi/inrrring Jrim ftirr l.'uii'miti/. Conoco Cftvmeafc OiMom Sltamck B. F. Goodrich StMflr Owntesi lintoet CattoiM WoUmlmot atocimm Rhono rwtono Qnv-tVa-Bm.--Ba.*cM waW Hurt Lorn WMorOhtnNt MomcdtoM tet SoNay Far Ian KsrahsQmaea aMSutTona &m*v ChMed Swwocno Otawai known by the authors n> have slmwn Home activity in t licensing in the PVC field. 'I'lic tabulation also includes companies licensing stripping methods for removal of unreurted monomer from the polymer product and licensing methods in inhibit the formation of settle on the surface* inside the polymerization reactor. These two areas have been ofj the highest research priority store the discovery of the, health I;/;vd jxwH'd by the vinyl chloride monomer' and the* need to reduce human ex|x>tture to very low level*. There are many patent* concerning PVC. produc tion. prinutlily groiqxrl in the ;tr<*a.\ ol |M>fyiiicri/atiim chemistry, reactor design, resin stripping and contmt of reactor fouling. While most licensors will haw patent protection in one or more of the processing steps, process and operating know-how i* usually rite most important psiri of licensor's package. RinuNcn 'llMlH'n. <mt(AMT, VWnl jn>1 AIIh-tI J.mrrv V.J 7 tu,. 1 a I'iT/i T,V__im__m_ll., W I...7 S. I'MrlM 1,'Vl.tVI Il'Mll. J.IM.I-A.l'lll, Stain* j'fat/jM, ).n I'l.T, i. 11*1 /Vaa.iaaa. I*iri, f. I'l '/Mvi*. W|H. Oil HrT*l, |. 17 Miltwr-Hi. K W, SUfk., M. 4ml Vr<it. (. | . /Mmailna, I'm*NUnti I'TI fc ,7J. . H.. -Ikv.kiwi*ni in P>( PnhMi. .ml in, i .-r, ) . Ol fin, H> M k- WtiKni. \ mi I >. isunn I,rw,i l7 I l*lJ I) JA Htlrm S.WJ.I U*l?H. kt. I, AtaiUni IVCvmcr Stirti*c. 21. t>. Ml? IfiT'ii ,(>fiM Pwcni *.jn7.il (1079). .ViwuKack. 1. X.HVC IHuwt. 9*. p 8*C(t*n7l. "Mi. M.T4**mk, M. andTin. |. M.w Hi., IS, ) 07 if)7l); V.* rim* , asm ').J*n>M. I. k, --t In- llfm ad j Hltlh.l W.4m Vuail.i .mi tnm ltam ,.VI.M aniiinni Pi4Miir<l/jM>Mi.* All 111 ?I| .ItMlll.il Mr..tuia ,J*A),Wlun ninit i,<i:>^KI I |*ITa* H.iK Ik. Ilium., taiinn. I hi'B. Ink w W.'W K. II, AmM Vmrr. I'l. (I*i77i `e.S. PMtm UHWIIIU7R). __ "t^nlWf. X. I'.*. P*m VISMtM <I<>7>1) 'jlkMrriulv. t. I ml l^rijn. I A . \ S r.l.i.l tl-lMli ,,I S Pmi'In MuVlHII'1711). I' S. r*MiriK I.111'i7i_ I S r^trni ).IW^1UH1 Hnnn. A r<mf lnkmi. I. h (I'iT'ii Alhtllln. I. F. ami NvtiHnf FimliM' l'imn>in. "Ik^fpia Ml OtM-jcan li Ik* I liirlAA I*. Itiirl. Rraiii.. nM4ki.nt/, K F,,. "iHllM.A'r l|a-.l Il4ii.tr* nh Hr. t La) Brjalamv'IW* fmm All* I'.I71 IUVr.tt(M> im Ctwtflkjl tHarnl FiiulnaMtH.* liwtuMrul Nnk IttM.rm il.marrh HU kllm.i S.'A ^(axMiiw SmibaB> ml |Yrf.imnrt l*-w<i>r Krtm l nh I Mnlluiv )kv M Mrf SMr(n," Mull. t.,, tVwr N. *7lil), Ii'hIi AunuJ At<4il Un.nq, Vm ^IScrn. Vurk (x. Mct namt * IImmr). Oi^mih* rt.a.mtr.," Rrusati li Rrtain Na. M.Anma . MlimW4im<r A--<uibi. fi (H Pmttitf.." IVifalr inniuni(4(a4i In.in (. | Knuu. t.i<Hirli4 hnwim. _Prlnjh (ax|i "~)V4n>nli hbaatr VanuIxiiiimiK Hmm*. " In.tn.'Mjl Km) I........ rl...mr,l mal In Miliant. Ana I. t`l7l ' FnlrtMl t.im<T. VrJ. .VI. N... I`M. F.rU.. >* I. !'?< 'lnbt.1 ka^pw.'T, V4. 11. No '(Ml, |)" Ill f'l" "-Iln- l,n|Mi ..I KtN>K*(4M. li.li>4.a-aH Imii# 4>hI < -|MjI Fa|iM|M1lrnl (m I'M Ham M-Waul. f4>m|KUUiMi 1*17*-1*BU*.' S)irnila<i I.i1-m4i.mv% )' and rStli* . I ttnrmitil Ann. Im 'Maun I<ula.it larinmah. In. .tin I laannu. I' < Wo. ** kavinnil.nrklHuMm*. !.*< I *tI. |a|ut., J<m.a I 1*1 , .1*. U Jir^a*. Mm In* u.an.ia ...ava (mill I'W.1 ILmdNaanm. l.iaKl. HanniiiiiiMa (.In-uik d to . I.ial a A" ' laluij )*> '.licwif. kitalum... Ilaa-Maala, Oukt. ItpaR. Khunr-hHitcnL. 47. rut alt Vilbtn. F-Mt .*>27 NruiU.nuf-Hrva- l Id* r.iBhH. wwimurHH ID-is, F O. Ban 2AI. 4fl*ki Ouninmut l, (.aetuum. Vt.imWu S.p A . (Jo rraiama liuu. Urm lkt*|am 2HI21 Mdan. ttul., A IOC.hi*niM>. |,u ..Auwtiint N... 4. *?*HUI Pjri. I* Drlrmr. Fraiar Mr.ia^. 1.1... *\.>uai'a Ahr.aal la.r l'\ C ami V M r l^mnml t *{<*m*t fir^rra V.d *V Si. I'l |.|. 11, | a i 50 MARCH 1980 HVOROCARBON PROCESSING AP00010794 \ r f rate* *ure vide >n. irecr cleri \. lallv sctijjh unl he for :ior ion :m. nn, ile- ii r! he iciciC Vinyl chloride "" STAUFFER CHEMICAL CO. Application: A process to produce vinyl chloride mono mer and 1,2-ethy)ene dichloride (EDO) from ethylene, chlorine and oxygen or air. Description: F.rhylene dichlorkle is produced in both the addition chlorination and uxychlorinalion sections or (he process. In addition chlorination, ethylene and chlorine arc reacted in the liquid phase to produce EDC; C*H + Ctr**C|HtCli + Heat The heat of reaction U used to distill the EDC pro duced in direct chlorination and oxychlorination scc. tiofi* of the plant, and recycle EDC from the cracking \ waion. This saves as much as 0.8 ton ofsteam per ton of ' VCM over conventional processes. Additional savings ` are realized by an equivalent cooling water reduction. In the oxychlorination section, ethylene, oxygen and HCl are reacted to produce KDC: C*H4 + 2 HCl + 'A Or-C*H4CU + H*Q This is a catalytic vapor phase reaction with the reac tion heat used to generate high pressure steam. Using trtygen rather than air reduces vent gas quantity by 95% and reduces energy and capital cost requirements. Vinvl chloride monomer (VCM) is produced by tracking purified EDC in a pyrolysis furnace: CsH<Clr-*CsHsCl + HCl After quenching, the products are separated into HCl, which is recycled to nxyrhlormaium. high purity VCM, and onreacted KDC which is recycled. Light and heavy ends from the process are inrinerated to HCl which is recovered. Alternatively, light and heavy ends can also be used as feedstocks for other chlorination processes. The aqueous effluent from the VCM plant is steam stripped and can be treated biologically. The VCM process can be "balanced" so that only VCM is produced, or the plant can be operated to produce HCl and/or EDC or to use "imported" EDC or HCl as raw materials. The process is automated for stable, safe operation with wide turn-down capability. In addition, the process is designed lor easy startup and shutdown. Manpower, capital, operating and maintenance costs are low. \ ) Commercial Installations:`I here arc M major unit* in operation or in construction with combined capacity of 8 billion pounds per year ofVCM and 19 billion pounds per year of EDC. Plant capacities range from 15 million to 1 billion Ibs/yr. of VCM. i i 1u Hydrocarbon Processing, November 1981 AP00010795 I - u^u.Id ^UUUl^LfulLlte / Vinyl chloride monomer... What you should know Till* survey of tftt VCM (vinyl chloride monomer) industry, commercial developments, chemistry, commerciel processes and new developments indicates that VCM can remain competitive with rising crude costs IW. McPherson, C. M. Starke and G. j. Fryer, Continental Oil Co., Ponca City, Okla. Development or the vinyl chloride monomer (VCM) industry has ln closely interrelated with the yvinyi chloride (PVC) industry. Effectively % percent the VCM production goes into the manufacture of FVC. Therefore, technological advances in one area have* jnificant impacts upon the other, producing a domino "ect. This, in conjunction with increasingly tighter rri- inUti regulations, has insured u continuing evolu- lion of VCM technology, ft is timely to review the state of the art. INDUSTRY PROFILE TIk* (-nnmieri'iiii ugniiiruiire of vinyl chloride monomer (VCM) cun be highlighted by the statistical ranking of the 19th largest chemical commodity in the United Suites. Turning our view upstream, we realize the significance VCM plays in wedding the |X>inx'licmical and chloroulkali industries. Pig. 1 schematically depicts the U.S. market integration surrounding VCM. Pondering the posture individual companies present to the market (Table 1), one can muse as to the motivational forces behind their respective business strategies. It holds that, if 96 percent of the VCM demand in this country is derived from PVC. then VCM'* future is incscu|wib)y tied to that of PVC's. Within ihc scope of known-unknowns, one basic fact attests to its longevity. On an energy-equivalent basis, PVC is one of the most energy-efficient construction materials available (Table 2). This follows even after weighing the socioeconomics of luedth stid environment. Looking at the two principal components of the PVC 1--Shows the U-S. market integration surrounding vinyl ehiohde industry. bon Processing March 1979 75 VINYL CHLORIDE MONOMER market, flexible or plaseirired products, of which fabric for automobile interiors and electrical wiring insulation are examples, and rigid products, including sewer and water pipes, electrical conduits and shoe soles, we find the rigid aggregate growing at 0 |jcrrrnt jkt year, as opjxMcd to the flexible area which is increasing at a rate of approximately 5 [xtreent. Within the past few yean, the rigid aggregate has surpassed the sire of the flexible market which ran hr witnessed hy the creeping seasonal resjxmse the VCM industry has to construction, the major end-use outlet for rigid products. Wood PVC Steel Aluminum Copper Pig. Relative energy content of various construction materials. TABLE 1--NamtpUl* capacities (MM pounds) Conoco.. Oiimend................. Deaf. Ethyl...................... 8. F. Goodrich......... ICI........................ tsenothem........... PPG.............. Shell., SUufler................ PVC 5SO 550 None Non* 175 1,100 None Nona None None 420 VCM Ethytone Chlorine 300 None Non* 700 550 None 1.000 Nose 2,500 2,250 4,800 8,400 300 None 500 MOO 350 2M 300 None 300 300 Nona Non* 900 None 7.900 I.MO 2,/OC 200 170 None 7M TABLE 2--Cemmarelal types of VCM lachnolopy I. Kith Parity AcabWns fndtttck II. Dilutaer Hied Gat Feedstock III. Balanced Ethylene Feedstock A. Air-Based Oxychlarinstton I. tanon-Based Oxychbrination Devalaper/Mctneor Dow................................................. Ethyl, ICI. Solray............................... B. F. Goodrich................................... Kureh*............................................. Mitsui Turin................................... Monsanto.......................................... PPG................................................. Rhone Poulenc................................... Steuffer.......................................... TokuyemeSoda.. . .. ToyoSods....................................... Koechst.......................................... ICI. .................... ............. Teeltnology Basle OttychlorintUon lit A III A ill A,9 n III B ill A.B ill B mA in A.B in A mA 1 1 . Over the course of the past ten \r.-ir. vc.Sf h.. come a major item of irilenmtinna) fnminrrr*-, wul United States the architect of this busmen*. Hetween anti 1977, eight jH'rtent of the VUM produced carried offshore, making it third only to styrene innnand toluene in the generation nf untie <hi]l,it- }\ 4|, cal commodity. This factor hsw tended to lill in tlw v; in demand brought about by construction. More sic candy, it has permitted the VCM industry to consistoperate near rapacity levels hy *x|x>rtng domestic pluses into tin? world urunu. I .nuking at tlx* global V market, we view forces at work changing the disjx* of trade. Beginning within the past five years. Eu as a whole has swung from a net imjtorter to rxp of VCM. With ten billion jxniuil* of VCM cap integrated to where then* is less than |.r> percent chantly derived PVC demand, producers arc now lenging U.S. material for <1 -share of the remaining v exports. Japan, whose five billion |>oimd VCM industry built to serve the whole of the Asian-South Pacific ma first in PVC and then later in VCM, has witnessmajor influx of U.S. and now F.um|x:nri monomer ducerx into their domnin. The present situation in the U.S. is manifestet cheaper feedstocks, vis-a-vis energy, ind an uriderva fiirrrncy that together are able in offset Icigistieul and remain competitive in the rmi'snining mar Eventually, merging energy purity and emission cost j throughs will leave U.S. VCM no more competitive that of any other nation. We rnm-lmlt, therefore, exports will continue as developing third world mat seek to establish a plastics industry iti advance of pt chemicals. Conversely, the major market imbalance the past hove g-nne hy tit** wayside, and market gm will romc from the domestic w-ctor. I'ig. 'I depict* longer u*rm outlook for VCM. In making our pmjerti we assume that the growing capital commitment rrqii to make VCM will not force' PVC In hecumr uiir<iiii| live with alternative products. Secondly. w<- foresee major technological innovnlioii on (lie hori/on dialer radically alter the economics of production. We do however, technological improvements of degree that reetinnnlly level out the iolliitionacy trend of plant 1 struction. Within the U.S.. we project a 700 to I. MM lb. per year grass roots plant will be required e\ two yean to meet demand. It is within ihe srn|>r <>f time frame that innovation will be tested. COMMERCIAL DEVELOPMENT VCM was first produced commercially in die ra 1900* via reaction of HC1 and acetylene derived fr calcium carbide, VCM usage in the manufacture of r thetic rubber accelerated dramatically during and al World War II. This increased demand prompted icarr for more economical hydrocarbon feedstocks. Acctyli was recovered from refuting steps, and new technok was deveto[>ed to produce acetylene specifically fr hydrnrarbon cracking. Ethylene became plentiful in the early 1950s. Dir chlorination processes to produce 1,2-dichlnroeth; (EDC) from chlorine and ethylene were developed 7fi March 1979 I Wor'icsriims I'nonss- AP000I0797 conjunction with EDC cracking technology to yield VCM. This process yielded byproduct HCI and did not pro liferate immediately, ex<-cpt in conjunction with acetyIcne-based technology which needed HCI to produce VCM. In the U.S., ethylene production from abundant sup plies of low cost LPC began to predominate. In Europe, ethylene prices also continued to drop, though not to the ume degree since higher priced naphtha and gas oil were the primary feedstocks. Therefore, while European producers continued to use acetylene-based VCM tech nology, American companies moved rapidly to ethylenebased technology. With the startup in 1938 of the first large scale oxychlorinatioa process to yield EDC from HCI and ethylene, a new era in VCM technology began. This '`balanced*' process allowed production of VCM from two commodity chemicals, chlorine and ethylene, without voluminous byproduct HCI. VINYL CHLORIDE CHEMISTRY Large scale production of vinyl chloride was first done by addition of hydrogen chloride to acetylene: Fig. *--U.S. vinyl chloride nameplate capacity versus demand. Hydmchlorination of acetylene Catalyst HC1+HC -- CH------------ - HC* = CHC1 (1) However, much lower costs for production of ethylene than acetySenr and the discovery that 1,2-tliohlorocthane (EDC) thermally decomposes to vinyl chloride in excel lent yield led to the following reaction sequence as the predominant manufacturing method for VCM: Direct chlorination of ethylene CH m CH, + Cl, -* CICH.CH.Ct EDC cracking to VCM Heat C1CH,CH,CI --------- CH, = CHC1 + IIC1 This method was esjwcially advantageous for those producers having a use for the HCI by product, particu larly so if acetylene were available; since then, a bal anced VCM process with no coproducts could be op erated. Later, the discovery that oxychlorinatioa technology could be applied to ethylene to give 1,2-dieMoroethane in high selectivity now allowed a completely balanced process based only on ethylene and chlorine as feedstocks. Oxychlorinatioa of ethylene Catalyst CH, - CH, + 2HC1 + 1/20, ---- C1CH,CH,C1 + HtO At present, about 92 percent of the vinyl chloride pr<v duced in the United Slates is from plants that use the balanced process based on ethylene via chlorination, cxycblorination, and thermal cracking of EDC.1 These three separate steps arc described in greater detail below. Additionally, the hydrochlorination of acetylene is also discussed below since plants using this chemistry are still in operation. Moreover, aomc recent crude oil cracking technology may narrow the cost gap between acetylene and ethylene with consequent revival of interest in VCM from acetylene. Some chemistry on direct preparation of VCM fmm.otlmnc is also briefly outlined, Direct chlorination of ethylene. Direct chlorination of ethylene to 1,2-dichloroethane is almost always conducted in a liquid phnsc reactor by intimately mixing ethylene and chlorine in liquid EDC. Ferric chloride, a highly efficient and selective catalyst for this reaction, is normally used in commercial processes. Amides, such as n,ndimethylformamide, have been reported to increase PV6' selectivity.* Oxygen, frequently present as an impurity in chlorine, likewise increases EDC selectivity in direct chlorination of ethylene by inhibition of free radical re actions that give 1,1,2-trirblnrocthnnr. Direct chlorination reactions may be run rich in either ethylene or chlorine, depending on the methods available to the plant for handling offgases from this reactor. Con version of the lean component is usually 100 |ercent, and selectivity to EDC is greater than 99 percent. 1,2-Dichloroethane, as it comes from the direct chlori nation reactor, is frequently nf sufficient purity for Track ing, except tltat it may contain ferric chloride, which would lead to rapid fouling of the cracking reactor. To avoid expensive purification of this already pure EDC. one may remove FeCl, by adsorption on activated carbon' or other solids.4 Alternately, one may operate the direct chlorinator at the boiling point of EDC, taking pure EDC overhead and using the heat of reaction to supply the heat for vaporisation.11, "T-" Oxychlorinalton of ethylene to EDC. Ethylene oxy- chlorination is normally conducted at temperatures of 225-325 C and at pressures of one to 15 atmospheres. Catalysts for this reaction almost always contain copper chloride and sodium or potassium chloride deposited on alumina or other suitable support. The detailed mech* Hydrocarbon Processing March 1979 77 VINYL CHLORIDE MONOMER anisrn of the catalyst's activity is not known, but it is recognized that cupric chloride is the active chlorinating agent. The cuprous chloride produced is rapidly recon* verted to CuClj under the reaction conditions, but the presence of some cuprous chloride is thought to be advantageous because it complexes with ethylene, bring ing it into contact with CuCL for a long enough time for chlorination to occur. The sodium or potassium chloride serves to increase EDC selectivity, mostly by inhibiting formation of ethyl chloride. Other catalyst components, such as rare earth metal chlorides, sulfate salts, ferric chloride and numerous other additives, have been dc* scribed in the patent literature. Good temperature rontml of the highly exothermic oxy reaction is a key clement in successful production of 1,2-dichloroethane. Temperatures higher than about 325 C lead to increased byproduct formation, mostly through increased dehydrochlorination of EDC to vinyl chloride followed by additional oxychlorination to give products having high levels of chlorine substitution. High temperatures also increase the amount of ethylene burned in carljon monoxide and carbon dioxide. Of equal im portance, high temperatures deactivate the catalyst by highly accelerated coking and consequent powdering * of the catalyst units and by increased sublimation of rop|wr, chloride away from the catalyst. Temperature control in fluidized bed reactors is main tained by the excellent intermixing of the catalyst par ticles and by use of internal cooling surfaces.1" Tempera ture control in fixed bed reactors is more di/TicuIt since "hot spots" tend to develop. To keep the hot spot temperature below below 325 C, yet get maximum utilization from the reactor, it is common practice to pack the reactor tube* with arrive catalyst arid inert diluent mixtures in proportion* of each so adjusted as to have low catalyst activity at the inlet, steadily in creasing to maximum activity nt the outlet. This grading of the catalyst activity flattens the temperature profile, allowing for good tcmjjcrature control with high produc tivity. For example, one patent* indicates the use of four zone* with 93 percent, 85 percent, 40 percent, and 0 percent, respectively, of the active catalyst pellets replaced by inert graphite. As an alternate to using inert materials in the eatalyst bed, catalysts, each with higher levels or CuC!| and consequently of increasing reactivity, are sometimes used. Fluid bed oxychlorination of ethylene, operated under good control, results in 94-07 percent ethylene ronversion, 95-97 percent HCl conversion, and EDC selectivi ty in the range of 94-96 percent. Fixed bed oxychlorinations are normally run with excess ethylene relative to HO, resulting in 93-97 percent ethylene conversion, 94-95 percent HC1 conversions, and EDC selectivity of 93*95 percent. These data do not include recovery of excess ethylene in subsequent reaction steps. Excess ethyl ene in vent gases from nxyrhlorination is normally con verted to EDC by direct chlorination with chlorine,,l,, although, if oxygen i* used rather than air, the excess ethylene may be recycled directly back to oxychlorination, Byproducts of ethylene oxyrhlorination are vinyl rhl* ride, ethyl chloride, 1,1-dichlorncthnne, vinyliderte chkv ride, nr- and tranj.1,2*dtrhlomethyIenes, trichloroethyl ene, chloroform, carbon tetrachloride, methyl chlorid*. methylene chloride, chloral ruiri high boiling enmjwtmAi All of these byproducts present problems in one way or another, such that their production needs to hr minimized to lower raw material costs, to lessen the difficulties tf preparing pure EDC, and to prevent fouling in the cracking reactor. Chloral, in particular, need* to be re moved since it jwlymcrivr.s readily in strong arid*#.' give solids which clog and foul operating lines and* controls. One must also take care to see that the feeds oxychlorination are pure. Normally, the only problem it with low levels (0.1 to 0.5 percent) of acetylene promt in the HCl from cracking of EDC. Acetylene in the Mj leads to the formation of considerable highly chlorinated !- byproducts and tan. Selective hydrogenation of this acetylene to ethylene and ethane is practiced by many companies!*^ Oxychlorination with oxygen instead of air. L> of oxygen instead of air for ethylene oxychlorination has received much attention.The outstanding benefia* from using oxygen are avoidance of expensive facilitis to recover EDC. ethylene and other chemical* from the large nitrogen vent g:i* stream: n large reduction in the quantity of olTgases that will probably need to be in* cinerated; and the ability to use ethylene as a diluent for oxychlorination, a procedure said to improve heat transfer in tubular reactors. Purification of EDC for cracking. Great care must t* taken to ensure that EDC used for cracking to vinyl chloride is of high purity, normally greater th:ut 99.5 per*; cent, since cracking is exceedingly susceptible to inhibition and fouling by trace amounts of impurities. Additionally,; the EDC must be bone dry to prevent excessive corrosion downstream of the cracker. For these purposes, one must consider removal from EDC of byproducts from three sources: EDC from direct chlorination, EDC from oxy* chlorination, and F.DC rcrnvcrcil from 11ir- cracking step (see below), EDC from direct chlorination is usually quite pure, greater than 99.5 percent: and, except for the FeCb present, it needs little further purification. As mentioned previously, ferric chloride may be removed by adsorption] on a solid, or the EDC may be distilled away from! FeCI* in a boiling reactor. Alternatively, the ferrir. chl>j ride may be removed by washing with water, usually in conjunction with oxy-EDC. \ 1,2-Dichioroethane from oxychlorination contains a variety nf impurities as listed previously. F.DC from this source is usually washed with water and then with caustic solution to remove chloral and other water extractabk impurities." Low boiling impurities, and water are takrn overhead in a first (light cnclsl distillation column, aid then pure dry EDC i* taken overhead in a second (heavy ends) column. EDC recovered from the cracking step contains an 7R March 1^70 ITviiK<>r:AK(i<>\ humii sstv.: APbdbio799 ppreciablc number of impurities, of which two. chlororcr>r and trirhlnrncthyicnr, are not readily removable * distillation, necessitating the use of other treatment*. ilorrrprcnr, if nnt altered by chemical treatment, concnimirs in the light ends rolnmn wliere it can |X>lymeri*e n olid nr rubbery materials which seriously foul and ipset tiiis column. Trirblomethylene forms an azeotrope vith HI )C, boiling very clow to EDC. If it is not removed n some way, it accumulates in the EDC, leading to re* luced cracking rates and increased fouling. Both impurii<*s may bn removed by subjecting the recycle EDC stream o rlilormulion prior t<i clistillulkm.1"'*3 Treatments with -ICl**'*41 and by hydrogenation*' have also been patented j methods for removal of ohloroprrne. 'racking of 1,2-dlehtaroathaita to vinyl ehtorldo. At ctniHTidun's in the range of 42.WKV* Ci, and ncur alnospheric pressure EDC undergoes clean thermal dehy* Irochlorination (cracking) to yield vinyl chloride and lydrogen cloride: Heat CtCHtCHiCI --------- - CH = CHC1 + HC1 The mechanism of this reaction has been extensively nvcstrgati:d*,` and shown to involve a sequence of free adica) intermediates. Use of pressure up to 25 to 30 atmospheres during tracking at temperatures of .'iOO-.WO C provides better lent transfer, rcdw^vl ,qmpin,nt siw and ensirr v-pnmJon of HCI from the product by fractional distillation. iDC conversion levels arc normally maintained in the angc of r0 to 6ft percent at residence times of 2 to 10 seconds, with selectivity of VCM ranging from 96 o 99" percent. Some byproducts generated during craek- 'g act as inhibitors to the free radical sequence so that r*'nsiri/' vv*rily leads to smaller tirul smaller inm-ast** n 1'ilXi conversion with rnrn's|>oiulingly increasing levels jf byproducts. Various materials, such as chlorine, bro* nine, or oxygen have been shown to be initiators for EDO cracking.** Recently, however, exclusion of oxygen s claimed to result in considerable reduction of fouling >n the cracker tube walls.** A rather spectacular claim tas been made that use of nitromethane as an initiator provides EDO convention lovrl* of up to 92.5 percent st 480 C/0 An important processing requirement in EDC cracking is rapid cooling or quenching of the re* tetion mixture. If cooling is done too slowly, substantial yield losses to heavy ends and tan rrnu!t.,l`**`1' Byproducts from the cracking reaction include acetyl* me, ethylene, methyl chloride, butadiene, vinyl acetylene, benzene, chlomprcne, vinylidene chloride, i.l-dichlorev ethane, chloroform, carbon tetrachloride, 1,1,1-trichloroethanc and other compounds. Most of these impurities remain in the unconverted EDC fraction and are re moved when this stream is distilled. Ethylene and acety lene codistill with the HCI and are thus routed hack to oxychlorination (after optional hydrogenation of the acetylene to ethylene). Methyl chloride and butadiene more or less codistill with the vinyl chloride, depending on the efficiency of the VCM fractional distillation sys tem. Addition of chlorine or carbon tetrachloride to the cracker feed is claimed to suppress methyl chloride for mation.33 Removal of butadiene, a contaminant which can interfere with )jo1yrnerization of VCM, has been done by treatment with chlorine,34 anhydrous HCI,33 or selective hydrogenation.-1" HCI addition to acatylan*. Recent development of a new crude oil cracking process53 using very high tem perature steam (2,000 C) as a heat transfer fluid pro duces substantia! yield of acetylene along with ethylene. Under some economic and geographic conditions, the use of this cracking process may be advantageous and may, thereby, provide acetylene for vinyl chloride pro duction. Typical conditions of hydrogen chloride addition to acetylene are total pressures on the order of five to 13 atmospheres, temperatures of 150 to 180* C, and use of a mercuric chloridc-on-carbon catalyst.31 With stoichio metric. quantities of reactants, essentially 100 percent con version is observed with VCM sclcctivities on the order of 98 percent. It is notable that ethylene does not react under these conditions, thereby allowing the use of mixed ethylene-acetylene streams as feeds. Ethylene, easily re covered from the vinyl chloride product by fractional distillation, it then chlorinated to yield 1,2-dichloroethane Other catalysts have been shown to be effective for addition of HCI to acetylene, but HgCI* is vastly supe rior.3* However, in addition to the general toxicity prob lems involved in working with mercury-containing substances,, HgCl, has appreciable volatility under the above reaction condition* lending to a need for consider able care and control in operation of the reactor. In fixed bed operations, HgCl, vaporizes from the hot spot of the realtors, condenses at cooler locations rimvnstrcarn and results in continuous movement of (be hot s|>oi downstream with eventual loss of catalytic activity. This loss is minimized by periodic reversal of flow through the mafrtnr. Ethan4)aMd vinyl chlorida processes. A number of patents dealing with chemistry for conversion of ethane to EDC and/or VCM have been published in recent years/*'44 Most of these reactions involve high tempera ture oxychlorinations, such as CuClj Catalyst CHsCH* + HCI + Of 350-450 C CH, * CMC! + 2H*0 However, these processes suffer from a number of dis advantages, the most tmjmrtanl of which include* low selectivity, low conversions and difficult operating con ditions such as CuCI, sublimation. One reaction system based on ethane, called the "TRANSCAT" process, has Iteen developed on a large pilot plant scale but is not yet in commercial practice.43'1' The chemistry of this process is a complex series of re actions, generally involving chlorination, dehydrochlori nation and oxyrhlorination. A mixture of'ethane, eth ylene, ethyl chloride, chlorine and HCI arc fed to a melt of cupric oxychloride and potassium chloride, yielding vinyl chloride, water, and cuprous chloride as the main products. The cuprous chloride-potassium chloride prod uct is taken to an oxidation reactor where the cupric oxychloride is regenerated by treatment with air. Vinyl chloride is separated from the organic product and puri- rfviiRnr\HiiMV I'unrKvMNi; Man h I07f> ?0 AP00010800 VINYL CHLORIDE MONOMER TABLE 3--Current VCM technology soure< Ord by fractional distillation. Removal of about 0.4 j>cr> cent butane plus butenes present in the main VCM cut represents a very difficult separation.1' Ethyl chloride and ethylene can be recycled into the feed. Chlorine value* tn the chlorinated byproducts can he recovered by incineration and then feeding these gases to the cuprous chloride oxidative regenerator reactor. DltpOtal of byproducts. Disposal of byproducts pre sents special problems for vinyl chloride manufacturing plants, since a variety of gaseous organic liquid, and acqueous streams must be handled, each with its own particular problems. Vent gas stream* from various units may contain small amounts of HCI, chlorine, ethylene, vinyl chloride, methane and carbon monoxide. These streams may some times be treated by scrubbing, chemical treatment, sorption, or other methods to recover some chemicals when economically justified. Otherwise, the common cleanup technique is either incineration or catalytic com bustion followed by recovery of HCI from the vent gases. Two organic byproduct streams arc produced. The light ends contain mainly ethyl chloride, eis- and irons* 1,2-dichloroethylene, chloroform and carbon tetrachloride. The heavy ends or "tars" contain mostly 1^1,2-trichloroethane, lesser concentrations of tetrachloroethanes, chlorinatrd butanes, chlorinated aromatics nml a large'number of other compounds present in small amounts. These streams are normally fractionated to recover useful comjnents, the other* incinerated to recover chlorine values cither as aqueous or anhydrous HCI. Process water streams are steam stripped to remove \ volatile organirs followed by nctitmlrrnlinn and then \treatment in an activated sludge system to remove non volatile organics in the water.4" COMMERCIAL PROCESSES Broadly speaking, there are three types of VCM pro cesses in commercial use today. These are categorized as acetylene, ethylene, or mixed gas based on feedstock re quirement. Within the "cthylene-typc" plants, further classification is desirable to distinguish the type of oxychlorination technology used, i.e., oxygen versus air feed stock (see Table 2). Over 90 percent of the world's listed 55 billion lb. per year VCM capacity currently is bused upon the balanced ethylene feedstock route. Of this, just under 90 percent uses air-based oxychlorination. How ever, of the projected worldwide plant startup* for the period 1979-1981, approximately 30 percent of the 7.7 billion lb*, per year VCM will be derived from oxygenbased oxychlorination. Other existing producers will un doubtedly he evaluating the conversion of present airbased plants to the use of oxygen during this time frame. Table 3 presents a summary of the technology sources in use today. It is thought that the technology currently licensed by 1CI and Solvny is similar to the process normally attributed to Ethyl. Therefore, all are listed under Ethyl's technology. It is interesting to note that several plants exist where the oxychlorination process TaehMlagy Stunt 1. B.F. Goodrich.... 2. Hoochsl/BFG*.... IS&ufftr/BFGV. 5. Ethyl,sotayJCI.. 6. Dow.................. 7. PPG.................. 8. ftbono-Peattfltc'.. 9. MtAsante........... 10. Tore Soda.......... 11. TokuyimsSod* 12. Mitsui Taatsu...... 13. Xufflha.............. 14. MIseallaMoos..... Existing Planned (1979- VCM. ------------ 1 V< Plants MV Lbt./Yr. Plants MM L 18 5,840 11 4,290 S 2360 19 6,390 14 5.000 7 3,040 S 1.700 5 1330 4 1,110 3 400 2 660 3 610 2 410 IB 2,400 4 2 7. 1 --_ 2 1 3 l.i 3 1. 1 l 1 _ 1 1 107 35,540 20 7. 4 faychtorination preens previtM by 8FG. of one licensor, B. F. Goodrich, has been combine direct chlorination and VCM technologies of othe Hnechst and StaufTcr. 'fable 4 presents a tabula worldwide VCM producers. This summary was pr from score* of literature sources, some of whicl contradictory. However, the authors have exercise best judgment in the absence of specific informatin the listed licensors. The reader is referred to Leonard,54 Gomi,10 or : for details of the acetylene and mixed gas routes tn The technology discussion herein will be limited halnnccd ethylene feedstock route used overwhcl today. Although each of the major technology license many patents, none lias complete c overage of ear of his process. As a result, the sequence and t operating steps tend to be very similar from pro< process. Thr basic difference* stern from lire oxyrlt lion technology and the iy|>e* of impurities apjrea the crude EDC. The licensor, therefore, provides know-how primarily, as opposed to patent positior result, published literature by major licensors is itandably very sparse and highly simplified, with exceptions. "5T-6">10 TYPICAL VCM PROCESS 'llie typical VCM process combines direct an. chlorination (oxy) processes to provide 1,2-dit ethane (EDC) feedstock for the EDC pyrolysis un Fig. 4). The direct chlorination process relative to oxy iacterized by low capital investment, low operatin and high purity product. However. HCI generate* thu EDO pyrolysis unit dictates the use of an ox With the current pyrolysis yield of approximate; mol VCM per mol EDC fed, anti the IICl yield o( per mol VCM, the oxy unit sire is set at approxi 0.5 mol EDC per mol of VCM desired. This s> direct chlorination unit size at approximately 0 EDC per mol VCM. The combined EDC streams are caustic treated lif) Vt.irili 1**7'* Ihimm MM;-.'. I'i AP00010801 T Flf. 4--Typicil vinyl chloride monomer bloek/flow diagrtm. Racier RMeter Reactor meetor manor VCM DC aurWmfw IDC Fl9' 5--Schematic of the Stauffer vinyl chloride monomer process. . .lIvr.HiKMuviN I'mir.i isim M.Ufli I 'I?'I VCM pyrMwMw HI APOOO10802 i TABLE 4--VCU plants--worldwide--(cont'd) wrator Hair AMc >Wl HtHIIKI M StIP SJneal {IfK Sir Caaaaala i*4*Owi Ji|<a auni nn CklM VCII ClilM Dank) Ami* lipiHH Cmk KanaRtfuCkt KaMRIflKfei Kaifeimt VCII Kuroha MtlwMtAKMtmanti MtluM TnM KiU*n y*.**fee*,t.td. Sanya Manamar sihwa Pitmfeamicili SvmIImw Sun ArmrCfeamfeii Tty* Gaaai Tty* SM KtPM Imm Padfe Kara* PMHe Libya CHOI Hum PtlMi *ftK iNLP IpHrtttttri Hydra PWM Setladad ParMHH CTOA. Palaad PMliMi'Ctltp Partdpi car Rmatto IMotliUI Impart Silt* Aglhauly SMIh Alrka Alow CipiaMM Chamteau Main ManiaMe lit Rodina WnkMf Otw MIRI KmiimM SwitMrUM IMU TllVIR Fmm Ptartfc TwMy Ptnuu Palkim Uafead IMw rltiali PwraRaaei liut T*<hma*i**an VlMIMil Palrhpllt THfMtovto Of*. Km*. tnd Haatilak* lad Dtw/ln* llUIH IrlndlN ParttMarpart CMUfl Caffe**) TiraMaj Imtru* Pitala Rayauta* Part# T*rrx CR*J nit HiniiPWt Cfeiki r*k**At Tthiwct rtiiun Haikimt Hiauu YakkrttM jsa* Chiba Mintfeini MlXMNni Thvtmi Ntthimi Twrrami TakKtMaa Tt*ayima City Yafckaithi Ya*-Su Ulun AMl Kammit* Pljirlta* Pajanlw MaManmadil PMaat Partmonn Wlseiawak Slna* Imtucu vtitaa Vilna f Saaatkgra \ Saatlkuri Tarrtfuj TiinfeM Haaferl Martaaall Math'* S!an*cw*d letdan MUui ladaa VMfeMaimii Akaaa-lMnir Mm tor. vtkiw HUDwau >a*e*nt OlarliMk Garkl Oorti Mlnfe VM*r* 7ima ti Tibia?* SUpic Panctn KIK Pracoaa U'O tf-%* r/vo kV-n /-/e C/A/B C/A/H m Af/A/- w m? ST E/A/l m f/A/l E/A/l f/A/l E/A/l f/A/O E/O/l F/aimW E/O/B A/-/- E/A/l f/A/l AH- C/A/1 C/O/l/ tf-f6 At/A/O V-/- mC/A/I f/A/l f/A/l %? I17t iiS I1M97S* 1M7 itrt i tit IMS mo mo !8S 1*4 \ta 11917709 ! m* ritnj (1900) m 197* 1971 0979) (I9*l> 197* (i9iei 1970 ((Ml) 1917 StSMfW ire ire RPC ppg Etkyt Maty Sir PPC Slwftir Taya Sadi Daaki Kacakg lapanau Caaa Staaffty Sirleivftir Karaka Maauata MtwJTaatw MtHMTaalw ST Takavaai Mi SUufftr Tchayiw* tad* Taya Saul Taya Ml Taya Mi Oaw OM* re, HoacftM Mireanual* Stuff** ire, MeMt Vukaa PPO oaar MitMi Toitw tCl, Safety Ttnnaea Haataaia Manual! Manual* Moaiaata Hliaaa PrafH Safety Oaw Slauffar 197) (iMi) mi 19^6 U 1977 (iiTij (1901) Shaft* Mitaai Taatw Safety ICI, SaArty ire Safety Safety saar-" PtMI ire.Haadirt Kurtk* ere, Hatcktl re Wuaa-Pral Similar Ota tAilAtar/ Caatrsotar PHtarWfeaalar Oft SaVti............ foiaea Safety ` Eulaea Sun*lam* Hitachi Samitaina Kanayalucht Chiyada ' Taya Eaenwnai JHallWy*lUCnimaannt Htlathi ' Pluar/DatVm Pratea r. URda, SautitMi lummat Bylaii Krakt ltdtar PftrKirben, Oivy-PoxRrfi* "lock Toye tnimaarint Hwmpfeilat A Ciistew C:iw!irt-*uuai 1 unmui *, cnr McKaa.'CTIP ' feck**, CTIP luil Haawkrty* and Ciauaw CTIP c. r. Brun C. r. Braun Spaiehim t.Uhda Chiyada f, Uhtla Pottar wfeatlar "*"T`'"Aja taui (oavorol) oi eanlradte(r (aoerlt, latfean kavt aurtiwd Diair bail iodpnanl ra/amnata atnola (feat ptmt > thaetfet ta b* m anpnaarnif aed/gr canttruebaa iiayt and nat Ml raaehad an aparakami ibtut LafaM: HrdrofffS?<! ^rtaok / Otylaai! / Pratau Emiylaaa cAJir fllalancttJ '0,,w" COliy 'D)irK( t&t. ill aa m 1U m m m III 91 770 mm ?it hi ui 71* 7IS IB IB IB U73t ut 1M 770 MO 770 ruism is 140 uo 17 >4S0> M m to us I70-US 3Mi UO 170 (HOT 2M am no u riS) III h S71 a 7) U $St Ml (IK) IIS (170) (U|) M.irrlt 1*17** T\ nt> i< Ml 1!' IS l'l< .1 I '.-l APOOO10803 VINYL CHLORIDE MONOMER table 4--VCM plant*--worldwicto OMriMr Uaitta Slitea ICI Arterial Cenoee Chtmxalt Oew CihWCorg. B.F. Cetarwfc Menerttn Shall Ckernwav Slwfir told** Chpmmi Ctargjg Pgi|< Oiimata Shamrock Algeria S.eennaellrrei ch Argentina CawQuinlca Eiaetroder Aeafealb IbOM * lihtckam Umkvrgia ILVM) sable lm ClfMi MlWIlt ViMlWM Patiaeolm Canteen Balfirb Tarteecewpitkl Canada Saar ShtvMiiln Chib Pctraeiitmtca Oaw/ENAP Chin Tadbeai Imparl Cara CtlaaBb Hlteeuimlec Columbun Cltrteglerikla Chemepetrel Ctomvkuggtar w. Kacka Finland Pctoma Or rnm AIZO Clunk EMC/DSM Diulac nat-tovlane SabK PCUK/Stoil Cbwnia bat Canutf JadnUrt Antigen Impart IWCmiay AltMtM Halt BASF Safety Byaarait Natal Naaahit . Untuck, AG Victor ia Caryl tallei Batata AUO Heart* StoHT Wengary Ctongktnpbi 1 v*rH KamhlMt Mb CIPItalt NiUaMl Orta be Chtrniglry Vta AOedee Ftlreehemieil Iria-Jipan l.v. to Slat* fetal (itctrachm>l Industrial Lac*lam Baba Beaut. Ueibeae UktCBarin. laMbna riaapan. Te*aa oitaacCrtrt. Tgu* WWtaMtaTliiMrta taba tay. (.abbeai Murtaa. Taint GefeerlCIty, Keatartv Gaiimar, lea-dan* Lata Chirtei. Landau LikeCBatbt, Laatuai suayaaHia. Faerie tea Deer tart, Taut Narce, Leaibau lan Beach, CaMem-e Caftmar. lemban* Piaeeeedat.itntdane Otar Park, Tins stitea Bahia Blanca Capiteinn. gBearnrataei laaia BBllai aca Bfearry Aeuara rainy Ttutednrle (amatpa.Jur Cklar Bahia Blaaca Camaetrl Bunn Dereye Serb*, Ontarie ran Stitiioiuii Varanaaa, Qeatac jhawimian, Quebec Conctptien Ceneapcien Peking Waritartea favaky, staaakb tanraa Cenlreudla Lt HMrt Ottrtcnbeim Jlltb Inara S>. Ataan SI. loan raaaui FetSvrMar Srtktpee WUhalauhawa Marl Lutawtrtaaen toalatari Luladarf Baabaci Keanart *SSS jJJSla BwjkaMM BaMak jJJ.Hta (CiteeMeei VtftMAgM Pamis Bertnto ereitto Kuiartttaka Madras Bembty Atadaa Bandar Haifa AUn Praeew c/Vi /VB E/A/0 e/vo a/ / 1/0/ B 1/0/ B E/O/B C/A/I/ E/A/B t/A/B C/A/B l/O/B t/A/B 1/0/8 C/V8 A/-/0 t/A/B l/A/t t/A/B r/A/B t/A/B 1/0/t C/A/B A/A/0 E/A/O E/A/B m Startup taar" ISU 1X1 isii mi IBM w im urc mr iixoi 1171 riwii tin 1171 ISU 1172 <li*i rllTV) riiH) 1X7 i XI mi mi \V4 1171 Ueenter BFQ Stertar Dew Dew Oew Ethyl Cthyt IFQ PPC PPG FPO Starter StauHir Starttr Stertar PPO Stowfar, IFQ Mlbei Tuba Daw ICI arc ICI Starttr FC Heart*. BFO Headist. BFG rihyl sahny, id BFC BFC PPC Oaw* Daw BFG BFC Daw V F. Caadrkb Slaular Safety r/A/B m At/-/* m m t/vo E/A/B W- t/A/B IX? ii>6 ix (iwdi ( tin ix 1*72 1X1 1171 1171 (117*) 0X1) i7i 1X7 1HI (II?*) 0X0) Btata'Piaml BIO, Hntktl Safety, Ettyl Bfb Haachii Hal* Starter BF4 HaedS Bn, Meertat tra bfb SSSS^* ICI Ethyl Sterttc/IPC Heiehat Heart* BFC, Heart* bfg, Heart* BFC Shall. BASF BFC Taya Seda Slirter Manunte Cnffewr/ CeecracMr V- M. Paraaae Fard, laaiaa a Omni................ B. M, Parnet Clear V. H. Pifmu V. M. Panaaa C. f, flriun J. it Firwat Brewaata Beat liimewn, B*d(t, Bream i Real Taya fa*nearing Chamiee BadyaV............. BASF Badiet SBaafdeaiayr MtKee. CTlP Badger Badfar/Praman TaehMp/TPl Bidgar Fibs Cagmeariig F. Uhda VCM . Capacity. m IBi/tr" seo 7 203 >.$ xo i I,DiO 100 ago MO 700 1 110 11.0001 1.000 rK) 110 *3 7 240 '1,1001 440 440 IX 220 f2Xl fl) r) 7 'MO IX m 7i r.. UMh.dP*araant. Vaact, Alpine CTIP tti'ana-pYaibf...... Rkant.Ptal a................. F.UMt Safeay C. F. Braee Uhta Badger, Ukba taK)utaa....... MerKuMb R. m. Pimeet .M.Pin#M Clear Cam prime-Ldrb AKio Cnginearfig UMa Badtar Badpar Badgar, lumniei Hitachi lunmut, Thyme 440 7H 211 441 44 rate) 4fll (Ml 700-710 IX 171 220 <440) 220 uo (UO) MO (ISO) (ISO) 79 3X IX 11 45 <3M) 141 21 r?20> 1IvriRfV:.\niMi-v | 'unr-i's- Vfnrrli 107rl APOOO10804 VINYL CHLORIDE MONOMER move HCi and certain clbrinated byproducts which otherwise would hinder fractionation or pyrolysis. The "dean" EDC is subjected to distillation steps in which water and other light eoinjxwients arc removed, as well as heavy components typically labeled tars. The dry product EDC, generally of 99.5 percent* or greater purity, is thermally cracked to yield HCI and VCM in an EDO carrier stream. Further distillation equipment separates EDC and HCI for recycle and yields product VCM for final treating. The typical VCM plant includes VCM treating, offgas treating, light ends/tan handling and waste treating facilities. It will also include incineration units for reclaiming waste chlorinated hydrocarbons from nfTgns or liquid streams.1 heat is removed by generation of steam on the shell r of each reactor. The final reactor 1*1110001 is cooled condense EDC, and thr offgas is conlacted/rcnclrd w chlorine to recover ethylene us additional EDC. T offgas stream is cooled versus cooling water and refr oration to further condense FIX1 before exiting the p' cess. Residual ethylene concentration in the vent gas reportedly as low as 10 ppm.*' Stauffer also offers an oxygen-based oxy process (F 6) in which the main reactor ofrgas, following rondere tinn of KDC, i* compressed anti recycled to the first o reactor. An excess of ethylene is used to maximize H convention and minimtw byproducts. ,\ small slipstita from the clhylctic-ricli recycle is purged to an cthyle recovery unit for control of inerts. Stauffer fechnofefy."*1 In the Stauffer direct chlori nation process (see Fig. 5), ethylene and chlorine ate reacted, in the liquid phase and under controlled condi tions, to yield a crude product which analyses 99.7 per cent EDC. The reactor product is then combined with the crude nxy EDC, washed and distilled to remove water, light ends and heavy ends. Pure EDC is preheated in the economizer of the pyrolysis furnace and then vaporised with steam. EDC vapor is tlicn heated to dissociation temperature in the furnace tubes to yield a mixture of vinyl chloride and hydrogen chloride. Conditions are controlled to maintain EDC conversion at 50 to 55 percent. Following n quench ;iiid condiMtsiili'Hi step, MCI, VCM, mid iiiir.riM'ktxl EDC are separated by distillation. Hydrogen chloride gas is sent to the oxy section. Unrearted EDC is recycled to purification. The oxy section combines recycle HCI with fresh ethyl ene and air in tubular fixed-bed catalytic reactors. The ethylene and air are fed in excess of stoichiometric re quirement* to assure high HCI conversion.*' fraction Ethyl Intagrated VCM procM." Gaseous chlorine s ethylene ure introduced into a direct chlorination react in which they combine to Term EDC. The very hi purity EDC product can be sent directly to (or, aft degassing, tain bypass) the EDC purification system ( Fig. 7). Air and gaseous ethylene and HCI are introduced m an oxychlorination reactor in which EDC is produced an elevated pressure and tem|xraturn in thr presence a fluidized catalyst. The reaction products arc neutralir and partially condensed to recover EIX1 which is first te to a drying column and then to the EDC purificati* system. A portion of thr vent gas. consisting prinwri of nitrogen and carbon dioxide is recycled to the react for added safety. The purified EDC stream which contains recyd EDC as well as the F.DC from direct and oxychlorir tion is vaporized and introduced into a furnace. At ]e; half of the EDC stream is cracked to HCI and VC! The reaction products are cooled rapidly, partially re densed, and then sent to the VCM purification v.*ir M.irHl | '17*1 | h l .no. .. ......... iv.... . AP00010805 Oirtci eMariMHan OiychlartoaUM ru g to ide to i tli 'he Rf. 7--Elhyl Corp.'s integrated vinyl chloride monomer proeass. iff* ro* : i HQ and VCM are separated hy fractional distillation to purification. Also, PPG uses three rather than two from the unconverted EDO and small amounts of by towen to obtain the HC1>VCM'EDC separation. iff- products. The EDC, containing the byproducts, it re cycled to the EDC purification system. - F. Goodrich technology.*'" Goodrich direct chlori- nation uses conventional water-cooled technology similar <V Cl UOtui Toalsu Chomicata technology." The MTC to that shown for Stauffer. The air-based, fluidized bed mi technology utilizes a boiling liquid process for the direct oxychlorination technology is similar to that shown for ne chlorination reaction. Reaction heat is dissipated with the Ethyl. Goodrich also utilises an absorber-stripper system gaseous EDC exit stream which is condensed externally on the oxy vent gas stream to minimize hydrocarbon ' and sent to purification. losses. Goodrich, in conjunction with Badger, Inc., offers id or The oxychlorination process is characterized by the use of oxygen feedstock and a fluidised bed reactor. The complete technology for waste treating of VCM plant effluent streams. fh reactor effluent gases are quench cooled with circulating Toyo Soda technology.*''*' This technology appears er EDC followed by caustic neutralization. The neutralized very similar to that offered by Stauffer. The principal ee gas is cooled to condense EDC and water. Uncondensed differences are in the use of an nbsorber.strippcr on the gases, primarily ethylene, arc recycled back to the oxy oxy vent gas effluent (as with Goodrich) and the de. to reactor. A small stream ts vented from the recycle gas hydration of crude EDC prior to purification. As with n to allow purging of inert!. EDC liquid is phase separated other oxychlorination processes, steam generation is used of from water and dehydrated before joining the EDC to remove reaction heat. d streams in the purification system. u K conventional EDC purification system splits crude Phono-Poutanc technology "-T` Rhone-Poulenc offers n EDC into light ends, heavy residue and pure EDC. The two processes, Chloe I and Chloe II- The former is of a y fetter is cracked to yield VCM which is purified in a special nature*' to yield concurrently significant quantities >r manner similar to that descrilicd in the Stauffer tech, of other chlorinated hydrocarbons such as trichloroethyl j nology (see Fig. 81- ene and tricMoroethane. The Chloe II process is "true" d | PPG t*Chn*4oy." The EDC production technology it ] appears very similar to that described for Mitsui Toatsu, I particularly in the use of oxygen feedstock and fluidized VCM technology using air-based, fluidized bed oxychlo rination in combination with boiling liquid direct chlo rination. bed reaction for oxychlorination. However, PPG does not Monsanto technology." This process appears very simi indicate use of a dehydrator to dry crude oxy EDC prior lar to that offered by Stauffer. Myiisocaruon Prcici'.xmnii Mnrrh if. AP00010806 VINYL CHLORIDE MONOMER PWH aMartnelton Maaetar OiyeMoHneeen Oimack column Caustic acruMar see pwwmum coe wrn*y* yen Pig. --Mitsui Toeteu uses these modifications in their vinyl chloride monomer process. DOW technology. Dow's technology has not -been pub licised. It has been used only by Dow and its foreign affiliates. NEW DEVELOPMENTS Although it U believed that several VCM producers currently uae boiling liquid reacton for direct chlorina tion, Stauffer has developed a unique application of this concept.4* Their approach, "High Temperature Chlorina tion," in effect uses the reactor as a reboiler for the con ventional EDO purification system (see Fig. 9) Purified EDC is withdrawn as a side stream from the tower, and any light components formed are removed overhead. Normal feed to the tower consists of treated F.DC from oxy and recycle. Small amounts of the normal heavy ends or tars arc purged from the hast* of the reactor. This ap plication eliminates approximately 100,000 lbs. per hour of 150 psig steam consumption for a one billion lb. per year VCM plant. A similar energy savings Is achieved in reduction of coaling water usage relative to a conven tional reactor and light ends tower. B. F. Goodrich" also offers a foiling liquid process in which the heat of reaction is utilized to purify all EDG processed in the purification train. Increased activity by EPA (U.S. Environmental Pro tection Agency) and state agencies in regulating hydro. Dkeet cMertnatten reactor and heavy ends column ends n. *--Stauffer high temperature chlorination and ethyleix dichloride purification schematic. I Mi M.inli VIIMIS Plow I ".IV AP000I0807 f sm .4 , oriwn emissions are likely to stimulate further new deiielopments, particularly in oxychlorination processes. [These will range from development of new oxygen-based tehnology to various add-on systems for cleaning up oxy [sent gas. The latter may include catalytic oxidation, in; oneration (of oxygen-based oxy vent gas), solvent absorp tion, combined refrigeration and absorption techniques [wd/er other combinations. Several companies not active ja VCM producers am involved in developing these ^idd-on systems. It is expected that companies will continue to devote Mfuiderable effort to the development of cracking prooters and inhibitors of side reactions in pyrolysis cherastry. Current cracking practice* limit EDC conversion - 50-60 percent. Considerable energy and cost savings [u!d be achieved through increased conversion levels jvithout concurrent losses of ET>C to undesirable side ; tactions. ;EPA regulation*. EPA's "Standard Support and Envi[mmental Impact Statement: Emission Standard for [Vinyl Chloride/'1 presented the following regulations: * Emissions from all point sources except oxyehlorina'fen would be reduced to 10 ppm VCM by volume a Emissions from the oxyehlorination reactor would be induced to 0.02 lb. VCM per 100 lbs. ElKi pmduet from the oxy process * Preventable relief valve discharges would not be ! permitted * Fugitive emission* would he minimized by requiring tftctaure of the emission sources and collection of the EPA estimated typical VCM plant emissions in 1974 is follows: Urt/VCM/ 100 Lbs. Source VCM Fugitive 0.1215 EDC Finishing Column 0.05 VCM Finishing Column 0.24 Oxy Process 0.0364 Process Water 0.0007 Total 0.4479 The regulations were predicated upon reduction of such (missions by 94 percent using best available technology. Compliance testing of these installations was begun in the last quarter of 1978. Additional EPA and state actions were initiated in mid-! 977 to reduce hydrocarbon emissions from VCM plants in non-attainment regions, i.e., much of the Gulf Coast. RDC production is reported to amount for 28 pfteem of ilu* hydrocarbon emissions in the southern Louisiana and East Texas AQCRs.** These actions were directed primarily against oxyehlorination vrnt gas from ur-based units. The amount of hydrocarbon reduction ought varies from region to region. No published guide lines are currently available to reference. The net effect of these various regulations has been to increase substantially the scope of add-on technology in VCM plants, such as: Installation of primary and redundant incineration facilities for VCM point (ex oxy) source nnd collected fugitive emissions Installation of HC1 scrubbing and neutralisation or recovery units in conjunction with the innticmiors Installation of closed proem sewers, collection sys tems and larger or redundant waste water strippers Replacement of single mechanical seals on pumps and agitators with double mrrh.mirnl seals. (In sonic cases, conventional pumps were replaced with canned or magnetic drive pumpa) Leak detection systems and portable monitors Enclosed sampling and analytical systems Vapor recovery systems for VCM Inading/imlnnclmg and equipment clearing. The EPA report estimated a maximum capital impact of $0.8 to $1.9 MM (1975 dollars) for a "model" 700 MM lh. per year VCM plant. Recent tost estimates in dicate the true impart for thrsr items is nearer $1.1 MM based oh 1978 dollars. Addition of hydrocarbon compli ance (proposed regulations) may add another $2-$5 MM. - F.PA also has proposed further reductions [n VCM emissions.** Under consideration at present arc regulations which will reduce allowable emissions from 10 ppm to 5 ppm for all point Sources, including oxy vmt Ran. F.PA further plans to prohibit emission increases within 8 kilometers of an existing source due to construction of a new emission source. This will effectively prevent expan sion of existing facilities or construction of new plants in the vicinity of existing plants. The proposed oxy vent gas regulation will alio dictate substantial capital expendi tures for add-on facilities and possibly the conversion of air-based to oxygen-based plants to facilitate incineration of tail gases. ECONOMICS Table 5 presents a 1981 manufacturing cost buildup for a typical 700 MM lh. per yrar grass roots VCM plant. Raw materials total 12 cents per lb. VCM or 54 percent of the required FOR plant selling price. Capitalrelated costa amount to 6.8 cents per lb. VCM or 32 percent. Utilities are only 6.7 percent of the total VCM cost. In perspective, the 1972-1973 reported VCM selling price** was only 4-5 cents per !b. By 1981, the capitalrelated unit costs alone will exceed this by 50 percent. The obvious major factor in VCM pricing is raw ma terials cost. Although chlorine, prim- is rxpwtrd to double Iwtwccn 1973 and 1981, the impact of ethylene price will be even greater (three cents per lb. versus 17 cents per lb.). The real Culprit, of course, is crude oil cosi. During the late 1960* and early 1970s, plants using inexpensive LNG feedstocks were significant contributors to the low cost U.S. ethylene supply picture. The energy crisis rap idly reversed the low cost feedstock trend. LNG scarcity IlYTIROCARTtn n Pnorr JUNO M.irrli 1*70 I i AP000I0808 VINYL CHLORIDE MONOMER dictated construction of naphtha and/or gas oil crackers for present and future ethylene production. This tied VCM prices irreversibly to rmdn oil prices through ethylene, fuel and power (particularly via its impact on chlorine). f*Y~ TABLE S--Estlmatad 1911 VCM manufacturing coat ACKNOWLEDGMENTS TWe authors gratefully achnwwlrdaa eontrlbmlsai by A. >. Strykaf, Jr.. of S" tau~ffer and' It. It. Watt of Ktatl and nenowaBt hr Mr eumpoatrt tu iw the iafi aiioo iwmlf M their LITERATURE CITED ' U.S. Environmeats! Protection Attney Report No. F.PA-4.VV2-7.5-(jn9, Research Triangle llik, N.U., (1975). Leach, M. (to Muaianto Chemical Co.) U.S. Patent },33Mf2 <1967). >*. F. Goodrich Co.. EritUt Patent 1,233,231 <|*7J), t Campbell, R. C., (ta Stauffer Chamkal Co.), UJS. Patent 4,000,205 (1976). Benedict, D., (to Union Carbide), U4. Patent 2,929,>52 (I960). Di Fiore, L, and Celmgoo, B., (to Son. Ital. RmIm), U.S. Patent 3,911,034 (l*ft|. 1 Ttao, U., (to Liimmni Cm.), VS. Patent 3,917,727 (t*73). Karts, B. XL.and Owliaa, A., <to Allied Chemical Co.), U.S, Patent 3.941368 (1976). Valean Material*, SHU* Patent 9*0.985 (1965). 'Van Antwerp, A. t., Hatpnag, ]. W., Stcrbcaa, R. C., and Kang, T. L-, (to B. F. Goodrich), UJ. Patent 3.488(1970). < Severing. T., On Stauffer Omniral Co.) U4. Patent 4.048m (F7>>. It Stauffer Chemical Co., Rrilish Patml 1,230,017 (1971). ' H. F. (rondrich Co., Belgium Patent 6811,415 (1966). Kiu, |{,, (w Mitsui Teauu Cbemkal). Japanese intent 4643367 (1971). Mlywahl, K.. (ta Mitsui Teatsu Chemical), Briilth Patent 1, 1*9,815 (1*70). / "Takahuhi. T.. (to Mltmi Toatsu Chemical), Japanese Patent 45*52408 U9J0). Mii'iii Toaiau Chemical, Japan** Paimt 46-3*117 fPWI). * PPC liwlinlrirs, PrearJi Patent 2,080,(did f T*171). " Mum. A. !.. dll Pit} Imhiilriet), Hrilidi IWnt 1.71*1,394 0971). * Ahlstrona, Jr., R. Cl., (ta Dow Chemical <J.), U.S. Patent 3,9fi*,*' (1976). "Strini, J. C.. and Castes, J. R., do RhonePrimil), U.S. Patent 9,959,386 (19781. About tho author* itOMSXT W. Mul'lUOtSON in products monoger, Continental Oil C*>, Houston. He is responsible for the worldwide profit performance of Conoeo's chlo rinated hydrocarbons and their xtmtegie development. Mr. McPherxm received his n.S, from Cornell University. Cuarum M, Stark* i* director of ex ploratory research, Continental Oil Co., Ponca City, Okie. Dr. Starke received hie IMS. from the University of Okla homa and his I`hJt. from Massachusetts Institute of Technology. GAXVtN J. FYA is supervising process engineer with Continental Oil Co., Ponca. City, Ohio. He i responsible for the nttper-vtnt'an of pracem deripn*, technical consulting and economics for chemicals processes. Mr. Fryar received his BJS. tn chemical engineering from (fc f/n*vritv f Mexico. 1. 700 MM NaadWwr Balloted VCM Flail. 2. INI Stamp . Crass Mata hwimmi - $140 MM, 4. Ufkt and heavy and* imfaeotod; memtaf HC! taM at mmittic stid to btssi ma aa iabaotatton mala. 1 Fdtooa osleast OCF rt4 U iftont toaever tatorsrt sksrteasnd profit. L UaH raiast an osaaraHtad and ara m soaeifn to a pedicular itsenier or lachntleff. " Smalley, E. W., Kara, B. and Bandyopadhyay, B., (tn Allied Qua* eal CorpO.Ui. Patent 4.06>4fl (1977). "Kaagaack Co,, fWitoafa Patent 1,206,070 (1972). Snlvny rl rir, Frmeh Patent I.HttStt IIVTI). * Kr-keter, A., (to Knanaack O.J, U..H. Patent 3,484,493 [tWJj. "Jaeklln. A. O., (tu ICt), flritUi Itotent 936.618 (1964). "PreelkA. (to HoccMt). Ooraaaa Patent i.217.694 (1973). * larioo, D. II. X., lTCwm. 5*>c., 148 (1949). "Young, D. P., (toB. P. Chcmieali, Ltd.), U.S. Patent 3,896,1*7 (t*T5). "Mitsui Chemical Industrie*, Japanese Patent 42-2292! li'JFi7|. " H. F. Goodrich. Rrithh Patent 938.824 (I'JUj. Knapaaek (>., U.ff. Puleal 3.47B;)IM (tto.')). n Moiisanlo ChrmiraMlu., Ilritidi Pitlriil 1,1118,371 (l%'l). M Keating, It. J.. (Ill Muuenlii (liemiml <.), (IJv. I'aieni $.125.(4)7 0*61). " danse, .. M . (it. M.miantu C>>niieal Co.), U.S. P.neni J.14J.7U,r [1K4I. "McDonald, D, W., (to Monsanle Chemical Co.), U4. ('stent It.llS.M ai, S,, Eighth World Petroleum Coogrem, Proceediart 4, 371 (1*71). MKrcna Chcnvral Industrin. Ilriildi Patent ')77,r>7n M`di4). Ilnliih 1`attM 1,(168.793 (IW), " ifartun, I). It., and Mugrian, M., J, .(<c Osin, /ad, ll^indon], fi*. 1) (1930); Petal, t., and WcitilicFi, P., flHv. Ckim. Art*.. 31. 743 (1949). "Gordon, K. D.. and Stsrfcs, C. M., (to Continental Oil Co.), U.S. PitMi 4.M6.m (1977). " Kuck. M. A., (In Siauffcr Chemlral 0>.), 1/.S, I'llrnl 3.mir,tie |I*M| * Whfilrin, N, J., (I.. Priiwei.t,, I'.I.miHil I... US. 1`dirv 3,551,5i (ID7(l|. "Kurts. It. K, lUalley, K. W.. Sommerman. W. K.. and Van Atu. J. *. (ta Allied Chemieal Coru.). U.S. Patent 3.987.119 11976). "Gordon, T. if., and Kttmmerle, II. F., (id Owani-Jflinoic. Inc.), U.S. Patent 4.042,639 (1972). "Ricgel, II., (to Lummut Co.), U.8. Patent 3,796,641 (1974). " Riegel, ll,, (to l.umn,i CV..), U.S. Patent f.V.T.WI (1971). " Khnel. II., (to l.ummut 0>.. U.S. Patent 3,937,744 tl97b|. "Sse, M, ()., (in l.nmmm On.). U.S. Patent VMt.ni'i 11976). "Tun, I)., (tu (,<imram (<.). U.S. IVlent (l`)7t,), "Tian. U., (to Lunimu* Co.), U.S. Patent XYJ2.4<4) (1976). > Riegel. II., (> Luautui Co.), U.S. Patent 3.933,286 (1976). " Taao, U., (to Loomus Go.). U.S. 1`atcnt 3,983,815 (1*76). Minot. J. D., CAem. ffag. Pregr., m, (8), 71 (1973). "Leonard, K. U.. P/njd sod Post Wmnrrt (i'srt 3), John Wilry |M Bona, Ino., New York, 1*71. " IWltisr, M.. PJajd CUmrU* aad PPG Man/Man, Noyrs Data Corp., 1978. "(iomi, S,, leses'i N*w PiaN CMerlde Prerrii Fnaraiii, VoT. 49, No. 11, November 1964. llyinttiU* * Raaenawaig, M. D,, /'Vlayt Process tits Wide Range ol Ry-Piodsctt,'' (RlMinr-Prngil), Ctvmlrd ffaginrrong. Vn) 78. No. 24. (h-i. Ifl, 10TI. "Kolirrl, J., and Itrrgier. A., "Rhone.l"i<sil New I'mxov' U* Ibe Msaw facture of Vinyl Chloride Monomer and Chlorinated Solvents /fan Zlkrk tn*," 164th A.C3. National Meeting, New Yark, Aug. 27>Sept. I. 1177. " Buckley, J. A., "Print*! Flow Shett/Vlayl Qiloridc via Direct QWsoiwa^aad uinchforisitia*," CAemaao/ agraeering, Vl. 73, No. 24, Nee. II. Frreale Cmnmomieatiaa From A. B, Stryker, Jr., Stauffer Chemical Ca Nev. 31. 1978. 11 Reich, INter, "Air or Oaygen for VCMf" ItyfSitet'Stn Prtetjntf Mart* 1976, pp. 1MI. Private rammiininiHno from 11. II. Wall. Ethyl Corn., Nov. 28, 1971. "Viayl Chloride, "Nydroearken ProeeaWlag. Nov., 1*73. " EPA-45(1/3-7346-C, Vol. 5. November. 1*74. Federal ffegirier. June 2. 1977. KPA-600/2-76^153, March. ]'J7G. * Keane, David P., et al, "Vinyl Chloride: (low. Where, Who--Future." j/ydrocerkoa Proeweiog, Fehrnnry. 1973. "Viayl Chtorith--Mitaui Toatsu Ghemieab." FFydrarsrkea Proeainag, No 1971. "Viayl Chloride--PPG Industries, Ine.," Nydroearksn PreasMtag, Nswober 1975. " "Vinyl Chloride--B. r. Goodrich Chemical Co., IlyiretttSe* PrtttuStt, November 1975. "Vlnri Chloride--Rhome-Peuleac S.A.," Hydfttt'btn Pnttnint, New ber 1973. ""Vinyl Chloride--(Monsanto Co.)/' Hydraeer6o* Prttstiim. NevnA* 1975. :i Private enanmuniritioas from J. S. ftrnuin It, K, Cm.ilrirh Chmiral C*. Jan. 3, 1979, JIM N-f.mh i"7'i list i1l > \ttll `S I'lini 1s|S AP00010809 *AVE MV LOVE picture? The firm could be sworn to secrecy on transfer price, if necessary. Authorize the auditors to present their findings to Congress and the public. Let's get profits out in the open! The alternative is a "peccable" audit by the government If higher gasoline prices or tax credits are required to build badly needed downstream refining capabilities, the public can be sold. And if the public is convinced, Congress will act The alternative is a shortage of refined products, especially unleaded gasoline. And when gasoline is not avail able for Americans1 love--their automobiles--politicians and oil companies alike will fsce their wrath. Coming noxt month Computer optimization and control, the June Special Report, is a subject teeming with potential. The reasons: greater conservation, operating flexibility and the advent of more reliable and less expensive computer control equipment. The report begins with the concepts of, the mathematical tech niques used, and places where on-line computer process optimisation really pays off for a variety of HPI processes. So existing plant* can evaluate the *poesible benefits of computer control, the second article provides a description of how to interface computers with existing plimts. Computer optimisation doesn't apply only to process areas. To exemplify this, the third article describes an optimal, energy saving, microprocessor based control system for centrifugal"compressors, June's issue is one you'll definitely want to keep on file. THE MAIL BOX Errata Several mistakes appeared in my ar ticle "Vinyl chloride monomer--What you should know," March 1979, pp. 7543. On page 77, right hand column, at the end of the twelfth line from the lop, "PVC" should be "KDC." On page 7ft, right hand column, second paragraph ending, the refer ence number "15" should be deleted. On page BO, Table 3, line B should read Khone-Poulenc. On page SB, Table f> npficar* cor rected below. Garvin J. Fryar Ponca City, Okla. table r furum imi veae wmrnciurhn ll* MMsrlsls IM M) UWfcmstrt.j.s..s.a..s..s................................_.....~........................... MUSK WU.l cast. W 14J IUNS 1.42 ISJ2 US ass MS 15.41 LSI 21.12 % itf S1.47 % 44 10.1 724 274 1064 i. m ms vcn nut i. mi stutM j. nn Mts Inraml $141 MM. i u*t ms hmrt *s6 imwnMr wwit hci nt it juristic mM M ft *k nn sa IACUMraMM tftfA 5. FiftoM sarcssl OCF rtM sf otars M *sr mtrwt efcsrsst anti mH i u.*.,)>** wwimr*miiii scnc* streatiruviRicrer sdwototy. | Hydrocarbon Procmsjno May 1979 for Huntington ALLOYS OIETK 6000$ OMSK* OF ALCAM ALUMINUM COW. OFFICES AND METAL SERVICE CENTERS: CALIFORNIA Sana Fe Seringa (M3) 6900011 COLORADO Denver (30313M-4141 CONNECTICUT Windsor (203) 66S-49M SCOWSIA AllSMa (404) 352-3070 (UINOtt cnieege (312) 625-3910 Prank!in Par* (31 4S5-7900 LOtlltliN* HMMU (504) 733-7310 MARYLAND aantmoraoen] 944-3000 laAsaaCHuacTTS Cambridge (917) 976-4600 Marlborough (617) 485-9400 (etcMOAN Midland (517) 631-2210 MISSOURI Kansas CUV <1) *71-35ie 8t. Lewis 13141427-1334 mewjereey Edison (301) 38*1700 NCWrORK Buffalo (Tie) 866-4900 Newrlbrtc City (313] 994-2800 OMO Cleveland (216) 624-6410 OKLAHOMA Tulsa (919) 999-2891 ORISON Portland (90S) 269 9991 PCNNSVUMMtA PWladelpMa (215) 464-3600 TENNE86BE Mewpna toot) 94a-34os TEXAS Beaumont (713) 842-6350 Dallas (214) 351-3371 Freeport (713) 233-6483 Houston (713) 7477--1n11io0 ly VMSSHINSTOH I2jJUKg*CTIU' ^settle (206) 768-0600 Spokane (509) 534-0686 6*M Matsis is gnr o* Metal Goad Circle 179 on ftaeder Ssrvict Can) 10) 1 I i AP0001081 ,1 ... ,ir,,r 2^ J >3 11*1IWI r5M"We)(cJ :'!..... .. ' ' '/ .WUI Alloy selection for VCM plants Oi(t aM MgfMtrt In aatadtng affeju for critical aqulpmant In dlract chlorination, oxycMorinatlan and EDC pyrofytit alloys for VCM plants in Europe, Japan and the USA. The data presented, within the limited scope of this article, arc intended only for guideline purposes, without any guarantee of performance. sactlena of vinyl chlcrlda monomar planta SELECTING ALLOYS 3). Nickel and high-nickel alloys are among the few me tallic materials having useful resistance to dry chlorine, hydrogen chloride and hydrochloric acid. These alloys are 4'. t4,l>. C. II. SeMllmotter, VDM Div., The Ore & Chrmir.nl not new to the chlor-alkali industry, bring mnrr or lew slumlord wlmium in cauxiic, brine and salt processing. Corp., Mmistnn (IHI.I We will, in turn, deal with each of the three chief corrosive environments encountered in VCM processes. Tabic 1 ;s Disign conditions, alloy selection and operating ex - provides a brief description or alloys commonly in use, perience are all part of the proprietary package offered their ASTM specifications and the tradenames under by various VCM process licensors. Therefore, of neces which they are known. HI sity, this article will he limited to the three separate operating step* that ail '`balanced ethylene feedstock*' Chloiiiw. Caseous chlorine at low tomj>cniturcs and in s (riant* have in common: direct chlorination, oxychlor- the absence of moisture is not severely corrosive and is i&ation and EDC pyrolysis. We will look into the cor commonly handled by carbon steel. Usually a more re rosive effects of chlorine and hydrogen chloride gases sistant material such as Alloy 400 is specified for critical u well as that of hydrochloric acid. Aspects of alloy se parts such as valve trim, instrumentation and orifice lection are based on published information of various plates in chlorine pipe lines. In contrast, wet rhlorinc processes, knowledge of corrosion and privileged infor is extremely corrosive on steel and nickel alloys, and re ,rt mation on actual tale* of large quantities of high-nickel quires Hastelloy* C or titanium. .At elevated temper I. |. (Editor's note. Allays 200, 201, 400, 600, 800 and 825 referred to in this article correspond to the VDM trade names ihown In Table 1 with which the author has the most recent experience in vinyl atures, corrosion rates on steel increase with temperature. The upper limit of usefulness is around 400 F when the protective effects of the corrosion products disap pear. Ferric chloride vaporises at 420 F. Fig. 1 provides a guide to the selection of various I. chloride monomer plants.) alloys in dry chlorine. The surface coating of chlorides tend to provide protection up to a temperature level TABLE 1 Alleys commonly usod In VCM plant* totartell "ffSSf* i Cr a* C uase VM CMentii F Wter MWter 1n44 HMt ra*IKtt Rfckd taM.......................... iwdrtoa Nlttil....... Alter m Alter rti n.i m lll-IU HI-HI VOM-MWUI2S0 MkhllTOO 111-1*3 111*1*3 VDM-NieMI 201 NieMi 701 mu MM* wif* lkttc*ptr Attey............. ........................... Alter MS 17 31 IS IU-113 US-US NICOMOS 400 MawlUO IkUl dv*al-- irw rlteyt *klf-CnrMmielfM Alter Alter (40 H IS 1 1*3-11) U3-IU NICSOrtRMO inwHlMO tdiMiwi-CMMigw Alls,. Alter MO 17 21 4* U3-407 1*3-4417 Nicsorutan ineeirr soo WkM-lrtn-CftiMiiuffi-M*lyMl'Wm-C*M*r Alter Alley US 47 11 3 73 N 1*3-473 1*3-423 MICRO*CR <2* incolDy *71 \t` meoftMS NlCftOfE* ii* rtfutrrM tntfiHMrtit* i HIMlUy it I brttmrfli t C*btt-S1*MM DWrsilft OntMfM Mttallrwfct *& (terminr. Msiwi, iw*ml. Incilif tr| nftittrid tratfmirfci al Tht IMfinitianii Nkil C HmsoriAannN P*otf.rrtn<; March **>70 tf, I V ) I AP000IO8I1 y ALLOY SELECTION t pi#. i--Chlorin#, Alley selection guide. FI#. 2--Hydrogen chloride. Alloy selection guide. it which melting, vaporiradoi? or decomposition removes such films. The corrosion rate appears proportional to the vapor pressure of the metal chlorides. The common design parameter for tubing, valve trim and internals is 0.003 in. per year (IPY) maximum corrosion rate, while for vessels and pipe an upper corrosion'rate of 0.020 IPY it frequently adopted with a corrosion allow* ance of Vi to x/+ in. Alloy 200 and Alloy 600 are the most commonly used alloys for reactors, coils and agitators in thr 500 to 1000 F range. Hyd*W" eWorld#. Dry HGI behaves in a similar man ner as chlorine, and carbon steel can suffice up to 450 F above which Alloy 200 is usually specified. Fig. 2 provides a guide to the selection of various alloys in dry HC1 gas and the design parameters of 0.003 IPY and 0.020 IPY upper corrosion rates for certain com ponents are shown. The presence of moiiturv does not appreciably increase corrosion rates above the dewpoint over that by dry gas. It must be pointed out that there are many variables, and even small amounts of addition agents to control catalyst activity, may exert an influence on the tenacity and vapor pressure of the protective corrosion scales. Therefore, absolute corrosion rates for different temper atures in chlorine and MCI systems arc difficult to pre dict. Even so it is felt that Figs. 1 and 2 are sufficiently accurate to serve as a guide. These figures should be checked against actual experience of each of the different VCM processes. The jicrformance of Alloy 200 in dry as well as wcl hydrogen chloride gas has been consistently good. In cyclic operating conditions, particularly in the presence of air or oxygen, Alloy 600 and Alloy 825 offer good all round resistance. It is prudent to assume that Types 304 and 316 stainless steels would be subject to chloride stress corrosion cracking conditions during shutdown in spite of various precautions that may be taken. 'i0 FI#. *--HydroeMorie aeld. Alloy selection guide (Buis: O.MO ' IPY max. corrosion rale). 1 Hydrochloric add. HC1 is w typu-iil reducing acid through its entire concentration range. Its strongly acidic character is harmful to steel. Alloy 200 and Alloy 400 be have similarly and find application at ambient temperature up to 20 percent concentration and at higher temper ature below 5 percent concentration. Fig. show* ijo corrosion lines for 0.020 IFY which is generally con sidered the upper design limit of an alloy selection. The graph immediately delineates the conditions suitable for handling with Alloy 400 and those where Hastelloy 8 is required. It should be noted that in most cases in which hydrochlorir acid is formed as a result of hydrolysis of chl-s f Vf.irrli 107fi .%11vihimi aum* AP00010812 r rides or chlorinated hydrocarbons such as EDC, the acid concentrations nrr less than .5 percent and both Alloy 200 und Alloy 400 ran withstand such conditions satis factorily at temperatures up to 400 F. Also in hydrogen chloride gaseous processes when cooling takes place near the dewpoint of HCI (that is j below 260 l1') it is prudent to assume thut because of the hygroscopic nature of the chloride on steel that HCI might condense with resultant high corrosion rates. In applications where 18-8 Cr-Ni stainless steel is used at ulcvulcd temperatures Midi coridcnsulinn run lead to chloride stress cracking. Alloy 800 is highly resistant to chloride stress corrosion cracking while Alloy 600 and Alloy 825 arc immune to this phenomena in this applica tion. Even though with xjxxiai precautions :t system can be kept dry during operation, adequate consideration ] must also be given to protect equipment during shut down and when starting up the unit. Purging with inert dry gas prior or during shutdown is helpful in sustain ing a dry atmosphere. Alternately one can maintain ) temperature* above the dewpoint. Also, sludges present in the system absorb chlorides and must be kept dry j otherwise aqueous MCI wid is formed from moisture pick-up. Alloy 400 is widely used for handling dilute HCI acid in waterwash solutions and for heat-exchanger tubes in the EDC purification. ftf. --Principal VCM proems slops- eacmemser |Hj To ewWtSHB" IOut CMemoeon reactor {DO4?) In EDC \ {^barton M V eonstrueden | Catalyst CyH4 Ct|--OH VCM PROCESS COMPARISON The typical VCM process combines direct chlorination and oxychlorinntibn processes to provide ethylene dichloride (EDC) feedstock to the pyrolysis unit. Because of the large amount of hydrogen chloride generated in the cracking of EDC, plant economies dictate the need for an oxy unit. Tim sequence find llirro principal njxrmlhig slr|xc are ipiitc similar from pnx-psx to proems and arc shown in Fig. 4. It is not die writer's intention to make a detailed analysis of the entire alloy selection for tin* various licensed processes, but instead, to rnmi merit on the basis of previously provided corrosion data, r on alloy considerations iiprilicitblr* in each of tin* three principal suqxt. PIq. S--Direct chlorination. Bract Chlorination. Ethylene is reacted with dry chlorine in lliu presence of a catalyst, to produce ethylene di- chloride. The catalyst usually is a ferric chloride. Liquid F.DC' is used as a reaction nuxlmm to insure intimate mixing. (2onvcntimi;d water cooling reuxivrs tlx* exo thermic heat of reaction. Temperatures are controlled at 130 F to 150 F. The RTX3 produced is usually treated for removal of ferric chloride carry-over. In a number of systems lately, the reaction is carried out at thr KDC Ixiiling jxiitit (approximately 230" F) taking pure EDO overhead and using thr heat required to Fig. --Oxychtorinallon, i vnporiw jwtn of the reactor content, to control the temperature. Fig. 5 shows the direct chlorination pro- on temperature and experience. Further, proper shut I cess. down procedures should be employed keeping the unit | It is necessary to insure u dry chlorine feedstock and dry or free from chlorine to prevent attack by wet to properly control the temperature by thorough mixing residual chlorine on the steel when the unit is not in of the reactants to prevent hot spots and runaway tem operation. peratures. With such ojxrrating control conditions, car bon steel can be safely used for the reactor and auxiliary Oxychtorinallon. Ethylene is reacted with dry hydrogen equipment. As can be seen from Fig. 1, a suitable chloride and oxygen in the presence of a catalyst to corrosion allowance may need to be applied depending produce EDC and water. The catalyst usually is im- i ISiin-i vsiNi. M.mli l'17'i i H AP00010813 alloy selection wham* sdc pwrtflwBw Cnmwtq rumM Oumn vcw pvrtflowen eotufnn n T--EDC pyrolyaii. pregnated with copper chloride. The oxyehlorination process can be carried out either in a fluid bed of catalyst or in tubular fixed bed catalytic reactors. Fur ther, the reaction of ethylene and HC1 can take place either with air or oxygen. Both process systems are shown in Fig. 6. Fluid bed. Reaction with a fluidized catalyst generally takes place at 435s F to 450 F and atmospheric pres sure. The reaction is highly exothermic and the tem perature is controlled by good intermixing of reactants and catalyst and by the use of internal cooling sur faces. A carbon steel reactor can be used. Referring to Fig. 2, we can expect conrxiinn rates in the order of 0.010 IPY which means a corrosion allowance of % in. nr 3/16 in. should he applied. For internal*, where corrosion can occur on both sides of the metal, where velocity effects may interfere with protective scales or wiwre n common allowance is not pmrticul or inter feres with the rate of heat transfer, a corrosion resistant alloy is commonly spedfled. Good service with Alloy 600 and Alloy 825 has been ob tained when used for the sparging equipment to intro duce the gases in the fluid bed, and for pipe, nozzle* and fittings inside the reactor. These alloys arc not susceptible to chloride stress corrosion cracking if hy drolysis takes piece during shutdown. Occasionally i problem is experienced from the combined effect of corrosion and erosion by the catalyst which remove the protective corrosion products on steel. Also, accu mulation of catalyst on the bottom of the reactor caa cause localized attack and should be avoided. Because of the carbon steel reactor construction, it is definitely necessary to control temperatures below 500 F. Proper shutdown procedures to prevent corrosion on steel by dilute hydrochloric acid, when the unit is not in oper ation, is of course mow desirable. Fixed bed- Reaction in a fixed-bed catalyst reactor is carried out at temperatures of 450 F to 600 1; and pressures from atmospheric up to 200 psi. The catalyst is contained in a number of vcrtiml tubes held in a tubesheet at top and bottom. Reaction heat is removed by the generation of steam on the shell side of the re actor. It is more difficult to control temperatures uni formly than in the fluidized lied, and occasional hot spots tend to develop on the tubes. As can be seen from Fig. 2 operating conditions re- | quire corrosion-resistant alloys. Both Alloy 200 and 600 ] are preferred alloy selection*. Usually Alloy 200 is used ) for the reactor tubes and has given excellent jwrfor- ' mancc. The tubesheets and heads of the reactor are clad j with nickel on steel. The reactor shell which is exposed f to water and steam is mndr of carbon steel. When 1 three or four reactors are required, nickel is also used ! for the interconnecting piping. Temperatures should be 1 controlled no higher than 600 I*" to prevent by-product ? formation and deactivation of the catalyst. Also, when j localized hot spots occur on the Alloy 200 tubes wr ? have observed intergmtiuiar embrittlement on some tubes. 5 Alloy 201, which has a low carbon level, is resistant to j Mich mtergmntilnr Rttnelc. ; Oxygen versus sir. The use of oxygen instead of air , in the oxyehlorination process improve* operating ef ficiency and product yield. It permits operation at a | lower Icmpcmtiire. Even though nllny sHcrtion is not utTer.ied by this, from a metals staudjtomt substantially ] less quantity of alloy is required as the equipment tends 1 to he more compact and smaller. Also capital savings ( are possible for not having to sc|x\retr HOC, ethylene and other chemicals from a large nitrogen vent gas stream and burning the vent gas. Lfirjh nickel alloys are often -`rio'c-d to locations .7? critical equipment Less 'Missive malerials are :e;: selected as soon ns iyt orocess conditionra EDC pyrolytls. Ethylene diehloride is vaporized and thermally rmckcd to vinyl chloride yielding vnlumma by-product hydrogen chloride. Cracking is carried out in a Bred furnace at a temperature of 900 F to 1000s F and at an elevated pressure of about 400 psi. The re action product* arc cooled rapidly, partially condensed and then sent to a VCM purification system where HCl and VCM are distilled from the unreacted EDC and small amounts of by-products. The cracking pro cess is susceptible to inhibition and fouling by trace amounts of impurities and therefore the crude EDC is purified and the light and heavy ends removed--both of which contain chlorinated waste products. Further, to prevent excessive hydrochloric and corrosion here and M.lCi It IM7M 1 (Mill. AUin i. 1'ii. i' 1 1 AP00010814 T downstream of llir rrrtrkrr, il is riftTMnry that ihr* crude KDC i* washed, nrulrolizcd with caustic, and About tho author dried, The pure EDO is preheated in an economizer and then vaporized in a steam heated ketde-type re* boiler, See Fig. 7 for the EDC pyrolysis process system. For alloy selection of the pyrolysis furnace tubing we C. K. "Dick" Sckiixmolloi i* rtnponaibla far th* technical marketing aeftvtHaJ for VDM hi th* USA. VMt j> a number of TietaUgeeelieehaft Group. Ho tt located fa Houeton, Ter&e. Be are guided by the fact that hydrogen chloride is formed in the thermal decomposition of EDC, that the furnace requires periodic decoking and that during shutdown rendition* hydrolysis of hydrogen rhlorirle is likely to wan. Alloy WK) is ihc ptrfrrirtl alloy sttlinu for this application and ha* consistently given good reliable per* formance. In recent years, Alloy 800 has been used with iuccess. It is prudent to assume that the 18*8 Cr-Ni type of stainless steels are subject to chloride stress corrosion cracking HHnw thr dew [mint nnd during shut* fore returning to th* United State* in fiareK If?* h had if year* urith in ternational Nickel fa th* USA, Aus tralia and Kurop* followed by a period of rnnmtjfffmm/ remanding in firttHurf*, lletffium. SehUim/tUvr hold* u degree m ehemfeai engineering. Ha is th* author of over too teehnieat papers dealing with oomoion and high temperature material problem* fa the proeees nufastriss, and ha* lectured at Stanford and the Univoraity of Californio. H* i* mil known to tho petroleum/peiroohomicitl induetry and serves an a number of technical committee*. down. One comjiany reported that they had used 5 Cr-'/t moly steel without excessive corrosion when in suring a bone-dry feed to the unit and taking the usual tpecia] shutdown and startup precautions of gas-blanket required to maintain such coating* and prevent damage at joints. ing and keeping thr unit dry. Alloy 600 is often speci TECHNOLOGY TRENDS fied for the cooling roil where halogen corrosion is the No major changes In the use of established materials of greatest. Also Alloy 600 solid or as clad plate has been construction is foreseen. Brief mention lias already been used for hatch polymcrizcr reactors in an aqueous me made of a trend towards higher temperature direct chlori dium in the presence of a catalyst and suspending agents, nation using the reactor as a reboiler to remove the heavy 0 d rd d n d ri n P s. producing PVC resin. * In the purification sysicm, corrosion can be very se vere because of the hydrolysis of hydrogen chloride nnd of various organi< chlorides. This can occur at temper- ttures of 260 F and downwards forming dilute hydro chloric acid. It is extremely difficult to insure a bonedry system, that is, '/cm free water and Mow 10 ppm of total dissolved water. Inadvertent moisture pick-up from sources such as leakage at water coolers and steam heaters and intrusion of humidity nt flanges and scats is always a possibility. Also entrained water is frequently carried along in the distillation process. Carbon steel ends. Also a substantial share of new plants appears to utilize oxygen instead of air in the oxychlorination step, This approach discharges fewer chlorinated hydrocarbon* with tin- vent gases into the atmosphere. For this reason and also to increase existing plant capacity, some VCM , plant oxychlorinators arc being converted from air to oxygen. Emission control standards are getting more severe, forcing the icrubbing and neutralization at HC1, use of waste water strippers to remove volatile organics, en closing and collecting emissions for incineration. Cor rosion in parts of this equipment can be very severe and '.(* in dry systems usually corrodes leu than 0.002 IPY while in the presence of water, dilute TTCl nod is. formed the rate of attack is not always predictable; some of the previously mentioned alloy selection guideline* do ir and corrosion rates arc approximately 0.010 to 0.160 apply. f- IPY depending on the temperature and concentration. Conclusion. For economic considerations, special high- a In the presence of 2 percent hydrochloric acid, carbon nickel alloys are often restricted to critical locations from 51 ited corrosion often exceeds one ineh per year. where one downgrades to Ion expensive materials a* Iv From Fig. 3 we ran see that Alloy 400, in most cases, soon as process conditions permit During the final de is Mould be the desired economical choice of alloy. It has sign, the limitations of each material selected must be e* been used in industry to resist a wide variety of chlori scrutinized as well as the means used to minimize cor IP nated hydrocarbons and solvents. Hastelloy B Ends use rosion by control of process conditions. For example, m for critical romjymcnt* such as valve trim and pump what will be done to control exothermic reactions, ve impellers. Corrosion of steel equipment tends to be the locity/impingement effects, moisture content, pH/acidic neatest in reboilers, the bottom section of distillation conditions, impurities, sludges and acid concentrations? id columns, bubble caps, plates, condensers, water separat Wn have the tools available to do this and particularly in us ion, valves, pumps and fittings. chlorine and HC1 processes we should not overlook proper ut F It should be pointed out that titanium Is not a suitable -election in dilute hydrochloric acid, because of its strong start-up and shutdown procedures, since severe corrosive attack may take place while the unit is not in operation. Ted reducing nature. Where titanium is used, it has provided ujeful life under mild cundilitms with a pH above 1.5. '"Vinyl Chloride," Hf4retrhS*nKT/R'lSuKtiNmC,H,i No--HtUr. 1975, pp. 214.216. re iC rw ce For economic reasons, various coatings and liners are frequently considered and used as an alternate to alloy. The following applications have hern reported: a gunrite Buehfrr, J- A.. `'Prnee** Fl-mk-tt Vinyl Ctilmririr via Direct <:iilnr>natii>i> 1 Kncdic(bl,iylf1m., ti"riAniirtitottri,MOCayMaeatnmitleei rXVaCfi.Mir,h" t|W, yNdtrancr-orhljernr 2/1`.re1m9CfIKlof. Mirth. 1926, ss. U-. 4 "Jte*dtanre of Nickel and Iliafc.Nickel Allow to Orrmifln In Hydro- rlilutic Arid, Clilutifle m4 Chlorine." (NCX) Cotttnio* I'jhv iC lined pretreatcr, glass-lined vessels, |lypropylenc-lined nrXecrliniitfaBounlletotinCCeErrBm-3io. nJ/u' lyHu1a9t7i3n.gteti Allow Twhaloal Srechure, Novoi> th -reel piping, Saran and Kynar-lined pipe, glass-coated tt, | wimps, various coatings in accumulators. Even though her, 1990. * ' SGIwelamhdielillsm.ieGaill.icoffn,.,CU".nEiMftf*e..",ctP!awaadtlrwMMitumuiniotunXt.etfm1anerC, oJrur"clyWn,hhm1a95tiSn.taPedtroochaebmouicialCoErnrovidroinng- nd e attractive from a first-cost standpoint, special care is mrnit," f'ktmiei/ Kntitutriat I'reerm, Oetolier. i960. B llVI'RtMiAKlUiK Pitt m ' | n.IN<; M.t I07`i `I", AP000I0815